CN116955429A

CN116955429A - Service recommendation method, device, equipment and storage medium

Info

Publication number: CN116955429A
Application number: CN202310193017.2A
Authority: CN
Inventors: 朱吉人
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2023-02-20
Filing date: 2023-02-20
Publication date: 2023-10-27

Abstract

The application discloses a service recommendation method, device, equipment and storage medium. The method comprises the following steps: acquiring index characteristic information of a plurality of service objects corresponding to a plurality of using characteristic indexes of a resource service and index characteristic information of service interaction indexes; carrying out service correlation analysis on index characteristic information of the plurality of service characteristic indexes and index characteristic information of the service interaction indexes to obtain service correlation information corresponding to the plurality of service characteristic indexes; determining respective service recommendation weights of a plurality of use characteristic indexes according to index weight representation data generated based on service related information; and based on the service recommendation weight, carrying out fusion processing on index feature information of a plurality of usage characteristic indexes corresponding to each service object to obtain service recommendation information of each service object for the resource service. By utilizing the scheme of the application, the accuracy of the service recommendation information of the service object can be improved, so that the service recommendation can be performed on the potential service object in a targeted manner.

Description

Service recommendation method, device, equipment and storage medium

Technical Field

The present application relates to the field of computer information processing technologies, and in particular, to a service recommendation method, apparatus, device, and storage medium.

Background

In the field of business recommendation, it is often necessary to mine out potential customers from a large number of business objects that are suitable for making business recommendations. If there are few business objects that have currently generated business interactions, a situation can occur in which the positive and negative samples are extremely unbalanced.

Experiments show that the potential objects cannot be well mined by adopting a machine learning method and a plurality of up-down sampling and threshold adjustment strategies under the condition, so that the effect of service recommendation and the success rate of subsequent service interaction are affected.

Disclosure of Invention

The application provides a service recommending method, a device, equipment and a storage medium, which can improve the accuracy of service recommending information of service objects, thereby pertinently recommending a plurality of service objects related to resource service and effectively improving the success rate of subsequent service interaction, and the technical scheme of the application is as follows:

in one aspect, a service recommendation method is provided, and the method includes:

acquiring resource usage characteristic information corresponding to each of a plurality of service objects and service interaction characteristic information corresponding to each of the plurality of service objects and aiming at the resource service, wherein the resource usage characteristic information comprises: index feature information of each of a plurality of usage characteristic indexes, the service interaction feature information including: index characteristic information of the business interaction index;

Carrying out service correlation analysis on index feature information of a plurality of service interaction indexes corresponding to the plurality of service objects and index feature information of service interaction indexes corresponding to the plurality of service objects to obtain service correlation information corresponding to the plurality of service interaction indexes, wherein the service correlation information corresponding to each service interaction index represents importance of the corresponding service interaction index;

generating index weight representation data based on the service related information corresponding to each of the plurality of usage characteristic indexes, wherein each parameter in the index weight representation data is used for representing the relative importance of any two usage characteristic indexes in the plurality of usage characteristic indexes to the service interaction index;

determining the service recommendation weight of each of the plurality of use characteristic indexes according to the index weight representation data;

based on the service recommendation weight, carrying out fusion processing on index feature information of a plurality of usage characteristic indexes corresponding to each service object to obtain service recommendation information of each service object for the resource service;

and recommending the resource service to the plurality of service objects based on the service recommendation information of each service object.

In another aspect, a service recommendation device is provided, the device includes:

the information acquisition module is used for acquiring resource usage characteristic information corresponding to each of a plurality of service objects and service interaction characteristic information corresponding to each of the plurality of service objects, wherein the resource usage characteristic information comprises: index feature information of each of a plurality of usage characteristic indexes, the service interaction feature information including: index characteristic information of the business interaction index;

the service correlation analysis module is used for carrying out service correlation analysis on index characteristic information of a plurality of service interaction indexes corresponding to the plurality of service objects and index characteristic information of service interaction indexes corresponding to the plurality of service objects to obtain service correlation information corresponding to the plurality of service interaction indexes, wherein the service correlation information corresponding to each service interaction index represents importance of the corresponding service interaction index to the service interaction index;

the index weight representation data generation module is used for generating index weight representation data based on the service related information corresponding to each of the plurality of use characteristic indexes, and each parameter in the index weight representation data is used for representing the relative importance of any two of the plurality of use characteristic indexes to the service interaction index;

The service recommendation weight determining module is used for determining the service recommendation weight of each of the plurality of use characteristic indexes according to the index weight representation data;

the service recommendation information module is used for carrying out fusion processing on index feature information of a plurality of use characteristic indexes corresponding to each service object based on the service recommendation weight to obtain service recommendation information of each service object for the resource service;

and the service recommendation module is used for recommending the resource service for the plurality of service objects based on the service recommendation information of each service object.

In another aspect, a service recommendation device is provided, the device including a processor and a memory, the memory storing at least one instruction or at least one program, the at least one instruction or the at least one program being loaded and executed by the processor to implement the service recommendation method according to the first aspect.

In another aspect, there is provided a computer readable storage medium having stored therein at least one instruction or at least one program, the at least one instruction or the at least one program being loaded and executed by a processor to implement the service recommendation method according to the first aspect.

In another aspect, a computer program product or computer program is provided, the computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the service recommendation method according to the first aspect.

The service recommendation method, the device, the equipment and the storage medium provided by the application have the following technical effects:

the application obtains the importance degree of the corresponding service characteristic index to the service interaction index by pre-determining the service characteristic indexes and the service interaction index aiming at the resource service, then obtains the index characteristic information of the service interaction index corresponding to the service object and the index characteristic information of the service interaction index corresponding to the service object, carries out service correlation analysis on the index characteristic information of the service interaction index corresponding to the service object and the index characteristic information of the service interaction index corresponding to the service object, obtains the service correlation information corresponding to the service characteristic index, the service correlation information corresponding to each service characteristic index represents the importance degree of the corresponding service characteristic index to the service interaction index, generates index characterization data for representing the relative importance degree of any two service characteristic indexes in the service characteristic indexes to the service interaction index based on the service correlation information corresponding to the service object, carries out fusion processing on the index characteristic information of the service recommendation weight corresponding to the service object, obtains the service correlation information of the service object corresponding to the service resource-oriented service, and has the service recommendation potential of the service interaction object, thereby effectively mining the service recommendation potential of the service interaction object.

Drawings

In order to more clearly illustrate the embodiments of the application or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the application, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of an application environment provided by an embodiment of the present application;

fig. 2 is a schematic flow chart of a service recommendation method according to an embodiment of the present application;

fig. 3 is a schematic flow chart of acquiring resource usage feature information corresponding to each of a plurality of service objects for a resource service according to an embodiment of the present application;

FIG. 4 is a schematic flow chart of performing subordinate clustering on a plurality of sub-objects based on a resource interaction relationship to obtain a plurality of sub-object groups according to the embodiment of the present application;

FIG. 5 is a schematic diagram of a label transfer provided by an embodiment of the present application;

FIG. 6 is a schematic flow chart of performing service correlation analysis on index feature information of a plurality of service interaction indexes corresponding to a plurality of service objects and index feature information of a plurality of service interaction indexes corresponding to a plurality of service objects to obtain service correlation information corresponding to each of the plurality of service interaction indexes according to an embodiment of the present application;

FIG. 7 is a flowchart illustrating another business correlation analysis performed on index feature information of a plurality of usage characteristic indexes corresponding to a plurality of business objects and index feature information of business interaction indexes corresponding to a plurality of business objects to obtain business correlation information corresponding to each of the plurality of usage characteristic indexes according to an embodiment of the present application;

FIG. 8 is a schematic flow chart of generating index weight characterization data based on service related information corresponding to each of a plurality of usage characteristic indexes according to an embodiment of the present application;

FIG. 9 is a schematic diagram of a business recommendation decision provided in an embodiment of the present application;

FIG. 10 is a schematic flow chart of a business recommendation method based on potential enterprise client mining provided by an embodiment of the application;

fig. 11 is a block diagram of a service recommendation device according to an embodiment of the present application;

fig. 12 is a schematic structural diagram of a service recommendation device according to an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

It is noted that the terms "comprises" and "comprising," and any variations thereof, in the description and claims of the present application and in the foregoing figures, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or server comprising a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.

It will be appreciated that in the specific embodiments of the present application, related data such as user information is involved, and when the above embodiments of the present application are applied to specific products or technologies, user permissions or consents need to be obtained, and the collection, use and processing of related data need to comply with related laws and regulations and standards of related countries and regions.

To facilitate an understanding of embodiments of the present application, several concepts will be briefly described as follows:

tag propagation algorithm: the main principle is to predict the label information of unlabeled nodes using the label information of labeled nodes. And establishing a relation complete graph model by utilizing the relation among the sample data, wherein in the graph, the nodes comprise marked nodes and unmarked nodes, the edges of the marked nodes represent weights between the two nodes, and the label information of the nodes is transmitted to other nodes according to the weights of the edges. In the embodiment of the application, the label information of the node represents the weight of the node. Tag data is just as a source, and non-tag data can be marked.

AHP: analytic Hierarchy Process, analytic hierarchy process. Is a systematic and hierarchical analysis method combining qualitative and quantitative analysis. The method is characterized in that on the basis of carrying out deep research on the essence, influence factors, internal relations and the like of the complex decision-making problem, the thinking process of the decision-making is mathematically carried out by using less quantitative information, thereby providing a simple decision-making method for the complex decision-making problem with multiple targets, multiple criteria or no structural characteristics. Is a model and method for making decisions on complex systems that are difficult to quantify completely.

Correlation coefficient: is a descriptive statistic that refers to the degree to which one variable corresponds to the change in another variable. Because of different study objects, there are various defining ways for the correlation coefficient, and the more commonly used correlation coefficients include: pearson correlation coefficient, kendall correlation coefficient, and the like.

Maximum eigenvalues of the matrix: let A be an n-th order matrix, if the number lambdaAnd an n-dimensional non-zero column vector x, such number λ being referred to as the eigenvalue of matrix a, the non-zero vector x being referred to as the eigenvector of a corresponding to the eigenvalue λ, holds the relation Ax = λx. The n-order matrix A has n eigenvalues (including the same eigenvalues), wherein the largest value is the maximum eigenvalue lambda _max ，Ax _max ＝λ _max x _max X is then _max For maximum eigenvalue lambda _max Is described.

The service recommendation method provided by the embodiment of the application can be applied to an application environment shown in fig. 1, wherein the application environment can comprise a client 10 and a server 20, and the client 10 and the server 20 can be directly or indirectly connected in a wired or wireless communication mode. The user may send a service recommendation request for a plurality of service objects to the server side 20 through the client side 10. The server 20 obtains respective index feature information of a plurality of service object corresponding service interaction indexes of the resource service and respective index feature information of a plurality of service object corresponding service interaction indexes of the resource service based on the service recommendation request, then performs service correlation analysis on the index feature information of the plurality of service object corresponding service interaction indexes and the index feature information of the plurality of service object corresponding service interaction indexes to obtain respective service correlation information of the plurality of service object corresponding service property indexes, the service correlation information corresponding to each service property index represents importance of the corresponding service property index to the service interaction indexes, then generates index weight representation data based on the respective service correlation information of the plurality of service property indexes, each parameter in the index weight representation data is used for representing the relative importance of any two service property indexes in the plurality of service interaction indexes, determines respective service recommendation weights of the plurality of service property indexes according to the index weight representation data, processes the characteristic information of the plurality of service property indexes corresponding to each service object based on the service recommendation weights, and feeds back the service information of the service object corresponding service object to the service recommendation end to obtain respective service recommendation information of the service object 10. The client 10 recommends a resource service to a plurality of service objects based on the service recommendation information of each of the plurality of service objects. It should be noted that fig. 1 is only an example, and the service recommendation method provided by the embodiment of the present application may be executed by a client or a server, or may be executed by both the client and the server, which is not limited in this aspect of the present application.

The client may be a smart phone, a computer (such as a desktop computer, a tablet computer, a notebook computer), a digital assistant, an intelligent voice interaction device (such as an intelligent sound box), an intelligent wearable device, or other type of physical device, or may be software running in the physical device, such as a computer program. The operating system corresponding to the client may be an Android system, an iOS system (a mobile operating system developed by apple corporation), a linux system (an operating system), a Microsoft Windows system (microsoft windows operating system), and the like.

The server side can be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, and can also be a cloud server for providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs (Content Delivery Network, content delivery networks), basic cloud computing services such as big data and artificial intelligent platforms and the like. Wherein the server may comprise a network communication unit, a processor, a memory, etc. The server side can provide background services for the corresponding client side.

The client 10 and the server 20 may be used to construct a system related to service recommendation, which may be a distributed system. Taking a distributed system as an example of a blockchain system, the blockchain system is formed by a plurality of nodes (any form of computing devices in an access network, such as servers and user terminals) and clients, a point-To-point (P2P, peer To Peer) network is formed between the nodes, and the P2P protocol is an application layer protocol running on top of a transmission control protocol (TCP, transmission Control Protocol) protocol. In a distributed system, any machine, such as a server, a terminal, may join to become a node, including a hardware layer, an intermediate layer, an operating system layer, and an application layer.

The functions of each node in the blockchain system include:

1) The routing, the node has basic functions for supporting communication between nodes.

Besides the routing function, the node can also have the following functions:

2) The application is used for being deployed in a block chain to realize specific service according to actual service requirements, recording data related to the realization function to form recorded data, carrying a digital signature in the recorded data to represent the source of task data, sending the recorded data to other nodes in the block chain system, and adding the recorded data into a temporary block when the source and the integrity of the recorded data are verified by the other nodes.

3) The blockchain comprises a series of blocks (blocks) which are connected with each other according to the generated sequence time, the new blocks are not removed once being added into the blockchain, and record data submitted by nodes in the blockchain system are recorded in the blocks.

In the following, a specific embodiment of a service recommendation method provided by the present application is described, and fig. 2 is a schematic flow chart of a service recommendation method provided by the embodiment of the present application, where the method operation steps as described in the embodiment or the flowchart are provided, but more or fewer operation steps may be included based on conventional or non-creative labor. The order of steps recited in the embodiments is merely one way of performing the order of steps and does not represent a unique order of execution. In actual system or product execution, the methods illustrated in the embodiments or figures may be performed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment). As shown in fig. 2, the method may include:

s201, obtaining the resource usage characteristic information corresponding to each of the plurality of service objects and the service interaction characteristic information corresponding to each of the plurality of service objects, wherein the resource usage characteristic information comprises: index feature information of each of the plurality of usage characteristic indexes, the service interaction feature information includes: index characteristic information of the business interaction index.

In this embodiment of the present disclosure, a service object may be a recipient of a service recommendation activity, and specifically, the service object may be a community including a plurality of subordinate sub-objects, for example, the service object may include, but is not limited to: enterprises, schools, etc.

In the embodiment of the present specification, the resource service may be a virtual resource service that needs to be recommended to the service object, for example, the resource service may include, but is not limited to: software product business, online course business, etc.

In a particular embodiment, the resource usage characteristic information for a resource service for any service object may characterize a statistical characteristic of resource usage for the resource service from sub-objects belonging to the corresponding service object. Taking a business object as an enterprise as an example, the resource usage characteristic information of the enterprise for the resource business can represent the resource usage conditions of a plurality of employees under the enterprise for the resource business.

In a specific embodiment, the resource usage characteristic information may include: the usage characteristic index may be an index for measuring usage of a service object related to a resource usage characteristic of the resource service, for example, when the resource service is an office software product service, the usage characteristic indexes may include: user liveness indicators (e.g., maximum number of simultaneous online people, average frequency of use and duration of users, etc.), teleconference initiation indicators (e.g., number of months initiated for teleconferences, average number of participants), shared document creation indicators (e.g., number of months created for shared documents, average number of participants), etc.; the index feature information of the usage feature index may be a specific feature value corresponding to the usage feature index of the service object, for example, the maximum online number of a certain enterprise is 532 people, the average use duration of users is 1h, and the month initiation number of the teleconference is 54.

In a specific embodiment, the service interaction characteristic information of any service object for the resource service may characterize the service interaction situation of the corresponding service object and the resource service. Specifically, the service interaction characteristic information may include: index feature information of the service interaction index, wherein the service interaction index may be a marketing descriptive index for measuring service interaction conditions, for example, the service interaction index may include: an interaction amount (e.g., an order amount), a number of interactions (e.g., a number of orders), etc.; the index feature information of the service interaction index may be a specific feature value corresponding to the service object under the service interaction index, for example, the order amount of a certain enterprise is 300 ten thousand, and the number of times of placing an order is 3.

In one embodiment, the plurality of business objects may include: a plurality of positive sample business objects and a plurality of negative sample business objects. Specifically, the positive sample service object may be a service object whose corresponding service interaction characteristic information is not zero, that is, the service object and the resource service have service interaction behaviors. In one possible scenario, a positive sample business object may be determined in conjunction with the actual scenario. For example, if the business recommendation is applied to a scenario of a software product recommendation, a first target number of business objects may be determined as a plurality of positive sample business objects from a plurality of business objects of a paid-for version of the product for which the software was purchased. Specifically, the negative sample service object may be a service object with zero corresponding service interaction characteristic information, that is, the service object and the resource service have no service interaction behavior. For example, if the service recommendation is applied to a scenario of a software product recommendation, the second target number of service objects may be determined as a plurality of negative-sample service objects from a plurality of service objects using only a free version product (trial version product) of the software. The size of each of the first target number and the second target number and the ratio of the two are not limited in the embodiments of the present disclosure.

In a specific embodiment, as shown in fig. 3, the obtaining the resource usage feature information for the resource service corresponding to each of the plurality of service objects may include:

s301, acquiring resource usage information corresponding to each of the plurality of sub-objects and resource interaction relation between the plurality of sub-objects and the resource service.

Specifically, the sub-object may be an auxiliary object under the service object, taking the service object as an example of an enterprise, the sub-object may be an enterprise employee, taking the service object as an example of a school, and the sub-object may be a teacher or student of the school.

Specifically, the resource usage information of the sub-object for the resource service may represent the resource usage situation of the corresponding sub-object for the resource service, taking the resource service as an office software product service as an example, where the resource usage information may include, but is not limited to: the duration of software usage, the number of participants in the teleconference, the number of edits to the shared document, etc.

In particular, the resource interaction relationship may be an interaction relationship generated by any action related to resource usage among the sub-objects, for example, the resource interaction relationship may include, but is not limited to: the interaction relation of participating in the same remote conference and the interaction relation of collaboratively editing the same shared document.

S302, performing subordinate clustering processing on a plurality of sub-objects based on the resource interaction relationship to obtain a plurality of sub-object groups, wherein at least one sub-object in each sub-object group belongs to the same target object, and the target object is any one of a plurality of service objects.

Specifically, when the resource interaction relationship is an interaction relationship of participating in the same teleconference, it may be considered that a plurality of sub-objects participating in the same teleconference belong to service objects corresponding to the teleconference; when the resource interaction relationship is an interaction relationship for cooperatively editing the same shared document, a plurality of sub-objects for cooperatively editing the same shared document can be considered as belonging to the business object corresponding to the shared document.

S303, determining at least one sub-object corresponding to each of the plurality of business objects according to the plurality of sub-object groups.

S304, extracting the use characteristics of the resource use information of at least one sub-object corresponding to each of the plurality of business objects to obtain the resource use characteristic information corresponding to each of the plurality of business objects.

Specifically, extracting the usage feature of the resource usage information of at least one sub-object corresponding to each of the plurality of service objects to obtain the resource usage feature information corresponding to each of the plurality of service objects may include: and carrying out statistical analysis on the resource use information of at least one sub-object corresponding to each service object to obtain the resource use information corresponding to each service object, and carrying out outlier processing and normalization processing on the resource use information corresponding to each service object to obtain the resource use characteristic information.

According to the embodiment, based on the resource interaction relation, the subordinate clustering processing is performed on the plurality of sub-objects, at least one sub-object corresponding to each of the plurality of service objects is determined, the use feature extraction is performed on the resource use information of at least one sub-object corresponding to each of the plurality of service objects, the resource use feature information corresponding to each of the plurality of service objects is obtained, the subordinate sub-objects of each of the plurality of service objects can be clustered accurately, and therefore accuracy of the resource use feature information corresponding to each of the service objects is improved.

In a specific embodiment, the plurality of sub-objects may include a plurality of first sub-objects and a plurality of second sub-objects, where the plurality of first sub-objects are sub-objects with subordinate object information marked therein, and the plurality of second sub-objects are sub-objects with no subordinate object information marked therein, and specifically, the subordinate object information may be a subordinate object tag of a corresponding sub-object, that is, the first sub-object may be a sub-object with a known subordinate object tag, and the second sub-object may be a sub-object with an unknown subordinate object tag.

In a specific embodiment, as shown in fig. 4, performing subordinate clustering on the plurality of sub-objects based on the resource interaction relationship to obtain a plurality of sub-object groups may include:

S401, a plurality of sub-objects are taken as nodes, resource interaction relations are taken as edges, and a sub-object relation network diagram is generated;

s402, carrying out tag propagation on a plurality of second sub-objects according to the sub-object relation network graph and the subobject information of the plurality of first sub-objects to obtain subobject information of the plurality of second sub-objects;

s403, performing subordinate clustering processing on the plurality of sub-objects according to the subordinate object information of the plurality of sub-objects to obtain a plurality of sub-object groups.

In the embodiment of the present disclosure, a node in the sub-object relationship network graph may represent a sub-object, and an edge in the sub-object relationship network graph may represent a resource interaction relationship between a plurality of sub-objects.

In the present embodiment, the subordinate object information of the plurality of second sub-objects may be determined from the sub-object relationship network diagram and the subordinate object information of the plurality of first sub-objects by a tag propagation algorithm (Label Propagation Algorithm, LPA), a tag transfer-based overlapping device discovery algorithm (Community Overlap PRopagation Algorithm, COPRA), or a community discovery algorithm (Speaker-listener Label Propagation Algorithm, SLPA). It should be noted that, the subordinate object information for determining the plurality of second sub-objects is not limited to the above algorithm, and the present invention is not limited thereto.

In a specific embodiment, the object label of the node corresponding to the first sub-object may be transferred to the object label of the node corresponding to the second sub-object in the sub-object relational network graph through a label propagation algorithm, so that the object label of the node corresponding to the second sub-object is determined under the condition that the convergence condition is satisfied. The convergence condition may be that the object label of the node corresponding to the second sub-object in the sub-object relationship network graph is updated by the label propagation algorithm until the object label of the node corresponding to the second sub-object is not changed any more. It may be understood that, since the number of update iterations required to wait for the object label to which the node corresponding to the second sub-object belongs to no longer change is more, the upper limit value of the number of update iterations may also be set, so that it is determined that the convergence condition is satisfied when the number of update iterations reaches the upper limit value. The convergence condition may be set adaptively according to the actual situation, and the present invention is not limited to this.

Optionally, in the process of performing propagation update on the object labels of the nodes corresponding to the second sub-objects through the label propagation algorithm, clustering of the nodes in the sub-object relationship network graph can be achieved, and under the condition that convergence conditions are met, the object labels of the nodes corresponding to the plurality of sub-objects in each cluster group are the same.

Optionally, the performing tag propagation on the plurality of second sub-objects according to the sub-object relationship network graph and the subobject information of the plurality of first sub-objects, and obtaining subobject information of the plurality of second sub-objects includes: constructing a probability transfer matrix and a labeling matrix according to each node of the sub-object relation network graph, wherein the probability transfer matrix is used for indicating the probability from one node to the other node in the sub-object relation network graph, the probability is determined according to the Euclidean distance between the nodes, and the labeling matrix is used for indicating the probability that one node in the sub-object relation network graph is labeled as each of a plurality of affiliated object labels; and repeatedly executing the step of updating the object label of each node in the sub-object relation network graph according to the product of the probability transfer matrix and the labeling matrix until the object label of each node in the sub-object relation network graph is unchanged.

Referring to fig. 5, fig. 5 is a schematic diagram of label delivery in an enterprise label labeling scenario according to an embodiment of the present application. Specifically, the sub-object relationship network graph includes a part of user nodes marked with enterprise labels and a part of user nodes of unknown enterprise labels, and edges in the sub-object relationship network graph can represent resource interaction relationships among users.

According to the embodiment, according to the resource interaction relationship among the sub-object nodes in the sub-object relationship network graph and the subobject information of the plurality of marked sub-objects, tag propagation is carried out on the plurality of unmarked sub-objects to obtain subobject information of the plurality of unmarked sub-objects, and finally subobject clustering is carried out on the plurality of sub-objects according to the subobject information of each sub-object to obtain a plurality of sub-object groups, so that the accuracy of subobject clustering can be improved on the basis of improving the accuracy of subobject marking.

S202, carrying out service correlation analysis on index feature information of a plurality of service characteristic indexes corresponding to a plurality of service objects and index feature information of service interaction indexes corresponding to a plurality of service objects to obtain service correlation information corresponding to the service characteristic indexes, wherein the service correlation information corresponding to each service characteristic index represents importance of the corresponding service characteristic index to the service interaction index.

In the embodiment of the present disclosure, the service related information corresponding to each usage characteristic index may be quantitative data for describing a correlation between the corresponding usage characteristic index and the service interaction index. Specifically, the service related information corresponding to each usage characteristic index may represent importance of the corresponding usage characteristic index to the service interaction index. Exemplary, service related information may include, but is not limited to: and the correlation coefficient, the feature importance degree characterization data and other quantitative data which can describe the correlation between the corresponding use characteristic index and the service interaction index.

In a specific embodiment, as shown in fig. 6, performing the service correlation analysis on the index feature information of the service interaction indexes corresponding to the service objects and the index feature information of the service interaction indexes corresponding to the service objects to obtain the service correlation information corresponding to the service interaction indexes respectively may include:

s601, determining a correlation coefficient between a target index and a business interaction index according to index characteristic information of the target index corresponding to a plurality of business objects and index characteristic information of the business interaction index corresponding to a plurality of business objects, wherein the target index is any one of a plurality of use characteristic indexes;

s602, the correlation coefficient corresponding to each of the plurality of usage characteristic indexes is used as the service correlation information corresponding to each of the plurality of usage characteristic indexes.

In practical application, the correlation coefficient is a descriptive statistic, which can be used for representing the degree of linear association between two variables, generally, the value range of the correlation coefficient is between 0 and 1, and the closer the absolute value of the correlation coefficient is to 1, the closer the relationship between the two variables is; the closer the absolute value of the correlation coefficient is to 0, the less closely the relationship between the two variables. Specifically, the correlation coefficient may include, but is not limited to: pearson correlation coefficient, kendel correlation coefficient, and the like.

In some embodiments, when the index feature information of the target index is a numerical variable, a pearson correlation coefficient between the target index and the service interaction index may be calculated, and the pearson correlation coefficient corresponding to the target index is used as the service correlation information corresponding to the target index.

Specifically, the calculation formula of the pearson correlation coefficient can be expressed as:

X＝{x _i |i∈[1,n]i is an integer }, Y = { Y _i |i∈[1,n]I is an integer }, where n is the number of business objects, x _i Index feature information indicating a target index corresponding to the ith business object, X indicates an index feature variable corresponding to the target index, y _i Index feature information representing the service interaction index corresponding to the ith service object, Y represents index feature variables corresponding to the service interaction index, ρ _(X,Y) The pearson correlation coefficient between X and Y is represented, cov (X, Y) is represented by the covariance between X and Y, σx is represented by the standard deviation of X, and σy is represented by the standard deviation of Y.

In some embodiments, when the index feature information of the target index is a category variable, a kendel correlation coefficient between the target index and the service interaction index may be calculated, and the kendel correlation coefficient corresponding to the target index is used as the service correlation information corresponding to the target index.

As can be seen from the above embodiments, the correlation coefficient between each usage characteristic index and the service interaction index is used as the respective service related information of each usage characteristic index, so that the objectivity of service related information assignment can be improved, and the interference of artificial subjective factors can be avoided.

In a specific embodiment, as shown in fig. 7, performing the service correlation analysis on the index feature information of the service interaction indexes corresponding to the service objects and the index feature information of the service interaction indexes corresponding to the service objects to obtain the service correlation information corresponding to the service interaction indexes respectively may include:

s701, taking index feature information of a plurality of service interaction indexes corresponding to each service object as a training sample, taking index feature information of the service interaction indexes corresponding to each service object as a labeling classification result corresponding to the training sample, and carrying out service interaction index prediction training on the tree model to be trained to obtain a trained tree model;

s702, analyzing the feature importance degree of the plurality of using characteristic indexes based on the trained tree model to obtain feature importance degree representation data corresponding to the plurality of using characteristic indexes;

s703, the feature importance characterizing data corresponding to each of the plurality of usage characteristic indexes is used as the service related information corresponding to each of the plurality of usage characteristic indexes.

Specifically, it should be noted that the model type of the tree model to be trained may be a random forest model, a gradient lifting decision tree model (GBDT) or an extreme gradient lifting tree model (XGBoost), and may also be any model based on constructing a decision tree.

Specifically, a decision tree of the tree model to be trained is constructed according to index characteristic information of a plurality of usage characteristic indexes corresponding to each service object. And mapping index characteristic information of a plurality of using characteristic indexes corresponding to each service object into a predicted service interaction index corresponding to each service object through the tree model to be trained. And carrying out iterative optimization on the tree model to be trained according to the predicted service interaction index and the marked service interaction index to obtain the segmentation points of each decision tree, and obtaining characteristic importance representing data of a plurality of using characteristic indexes according to the determined non-purity of the segmentation points.

Specifically, the feature importance degree characterization data of each usage characteristic index may characterize the importance of the corresponding usage characteristic index to the business interaction index prediction.

In some examples, the feature importance of each usage characteristic index may be evaluated by using the relative order in which each usage characteristic index is used as a decision node in the decision tree. The usage characteristic index used at the top of the decision tree will contribute to the final predictive decision for more samples. Thus, the feature importance of each usage characteristic index can be evaluated by the proportion of samples that each usage characteristic index contributes to the final prediction. In the example of the XGBoost tree model, a score corresponding to each usage characteristic index may be obtained by a feature importance score feature_importances_and used as feature importance characterization data.

According to the embodiment, based on the trained tree model, feature importance analysis is carried out on the plurality of usage characteristic indexes to obtain feature importance representation data corresponding to the usage characteristic indexes, and the feature importance representation data corresponding to each usage characteristic index is used as service related information of each usage characteristic index, so that objectivity of service related information assignment can be improved, and interference of artificial subjective factors is avoided.

S203, generating index weight representation data based on the service related information corresponding to each of the plurality of usage characteristic indexes, wherein each parameter in the index weight representation data is used for representing the relative importance of any two usage characteristic indexes in the plurality of usage characteristic indexes to the service interaction index.

In the embodiment of the present disclosure, the service interaction decision between the service object and the resource service is affected by the multiple usage characteristic indexes to different extents, that is, the importance of the multiple usage characteristic indexes to the service interaction index is different, and accordingly, the index weight representation data may represent the relative importance between two indexes, where the two indexes are all usage characteristic indexes in the multiple usage characteristic indexes.

In a specific embodiment, the index weight characterization data may include a plurality of parameters, and each parameter may characterize a relative importance between two indexes corresponding to the parameter in the plurality of usage characteristics indexes.

In a specific embodiment, the metric weight characterization data may include: the matrices are compared in pairs.

The pair comparison matrix is a quantity basis in the analytic hierarchy process, taking N use characteristic indexes as a plurality of use characteristic indexes as examples, and in the embodiment of the application, n×n pair comparison matrices can be formed for the N use characteristic indexes. For any one of the use characteristic indexes F in the use characteristic index sequence F formed by N of the above use characteristic indexes _i (1.ltoreq.i.ltoreq.N) and a usage characteristic index F _j (1. Ltoreq.j. Ltoreq.N), using the characteristic index F _i Compared with the usage characteristic index F _j The relative importance of the service interaction index is the parameter M in the paired comparison matrix M _ij Is a value of (2).

In a specific embodiment, as shown in fig. 8, the generating the index weight characterizing data based on the service related information corresponding to each of the plurality of usage characteristic indexes may include:

s801, obtaining a use characteristic index sequence according to a plurality of use characteristic indexes.

Taking a user liveness index, a teleconference initiation index and a shared document creation index as examples, and the obtained use characteristic index sequence F is taken as an example, wherein F ₁ Indicating user liveness index, F ₂ Representing teleconference initiation indicators, F ₃ Representing shared document creation metrics.

S802, any one first index and any one second index are obtained, wherein the first index and the second index are the use characteristic indexes in the use characteristic index sequence.

S803, determining a target parameter according to the service related information corresponding to the first index and the service related information corresponding to the second index, wherein the target parameter represents the relative importance of the first index to the service interaction index compared with the second index.

In a specific embodiment, the determining the target parameter according to the service related information corresponding to the first indicator and the service related information corresponding to the second indicator may include: and taking the ratio of the service related information corresponding to the first index and the service related information corresponding to the second index as a target parameter.

Exemplary, if the first index is F ₁ The second index is F ₂ The service related information corresponding to the first index is f ₁ The service related information corresponding to the second index is f ₂ Corresponding target parameters

S804, determining the position of the target parameter in the index weight representation data according to the position of the first index in the use characteristic index sequence and the position of the second index in the use characteristic index sequence.

Exemplary, if the first index is F ₁ The second index is also F ₁ The corresponding target parameter is m ₁₁ In the case that the index weight characterization data is a pair-wise comparison matrix, the target parameter m ₁₁ Located in the first row and first column of the pair of comparison matrices M.

S805, generating index weight representation data according to the target parameters and the positions of the target parameters in the index weight representation data.

By user liveness index F ₁ Teleconference initiation index F ₂ Shared document creation index F ₃ The obtained usage characteristic index sequence F= { F ₁ ，F ₂ ，F ₃ For example, in the case where the index weight characterization data is a pair-wise comparison matrix, the pair-wise comparison matrix M may be represented as:

in a specific embodiment, the service related information corresponding to each of the plurality of usage characteristic indexes may include: the plurality of correlation coefficients corresponding to the usage characteristic indexes respectively, the index weight characterization data may include: the pair-wise comparing the matrices, correspondingly, generating the index weight characterization data based on the service related information corresponding to each of the plurality of usage characteristic indexes may include:

a pair-wise comparison matrix is generated based on the correlation coefficients corresponding to the respective plurality of usage characteristic indexes.

Specifically, the refinement step of generating the pair-wise comparison matrix based on the correlation coefficients corresponding to the plurality of usage characteristic indexes may refer to the refinement step of the index weight characterization data generation process of S801 to S805, which is not described herein.

In a specific embodiment, the service related information corresponding to each of the plurality of usage characteristic indexes may include: the plurality of feature importance characterization data corresponding to the characteristic indexes may include: the pair comparison matrix, correspondingly, generating the index weight representation data based on the feature importance representation data corresponding to each of the plurality of usage characteristic indexes may include:

and generating a pair comparison matrix based on the characteristic importance degree characterization data corresponding to each of the plurality of using characteristic indexes.

Specifically, the refinement step of generating the pair-wise comparison matrix based on the feature importance characterizing data corresponding to each of the plurality of usage characteristic indexes may refer to the refinement step of the index weight characterizing data generating process of S801 to S805, which is not described herein.

In the related art, the construction of the comparison matrix is a more important step of the analytic hierarchy process. The analytic hierarchy process is a decision making process of decomposing elements related to decision into layers such as a target layer, a criterion layer, a scheme layer and the like, and carrying out qualitative and quantitative analysis on the basis of the layers, and is summarized by calculating the weight of each index, and obtaining the total score and the sequence of each object through weighting. Referring to fig. 9, fig. 9 is a logic schematic diagram of a service recommendation decision provided in an embodiment of the present application, firstly, an analysis model of a three-layer structure is built based on a hierarchical analysis idea, Q service objects required to perform service recommendation analysis are determined in a scheme layer, N usage characteristic indexes for calculating service recommendation information are determined in a criterion layer, and service recommendation information (service recommendation degree) of each service object is taken as an analysis target in a target layer. And determining a pair comparison matrix according to the relative importance of N using characteristic indexes in the criterion layer, and obtaining weight vectors corresponding to the pair comparison matrix, so as to obtain service recommendation weights corresponding to the N using characteristic indexes, and calculating service recommendation information (service recommendation degree) of each service object according to the service recommendation weights.

In the prior art, when an analytic hierarchy process is applied to construct a pair of comparison matrixes, only subjective judgment of an evaluator (namely expert) is generally used for assigning values, and the obtained pair of comparison matrixes is biased to subjective, so that the accuracy of service recommendation weights corresponding to the use characteristic indexes is affected; in addition, in order to prevent logic errors in the assignment process of the pair comparison matrix, consistency check needs to be performed on the assigned pair comparison matrix.

Therefore, the present application proposes a hierarchical analysis method based on service related information to solve the above-mentioned problems in the prior art. Specifically, in the application, the correlation between the index of the criterion layer and the analysis target of the target layer can be described by adopting the service related information (such as the correlation coefficient and the characteristic importance representing data), and the correlation scale (namely the parameters in the paired comparison matrix) generated by 1-9 and the reciprocal thereof in the existing analytic hierarchy process is replaced by adopting the ratio between the service related information, so that on one hand, the data support is increased, the objectivity of assignment of the paired comparison matrix is improved, and on the other hand, the accuracy of the finally obtained weight corresponding to each use characteristic index is improved, and on the other hand, the paired comparison matrix constructed by the analytic hierarchy based on the service related information is necessarily a consistency matrix, thereby avoiding consistency check of the paired comparison matrix and improving the efficiency of service recommendation analysis.

The following demonstrates that the pair-wise comparison matrix constructed by the above-described hierarchical analysis based on business-related information is necessarily a consistency matrix:

because, in the present application, the parameters of the pair-wise comparison matrix are determined based on the ratio between the respective service-related information of any two of the N usage characteristic indexes. Schematically, the pair-wise comparison matrix m= (M _ij ) _N×N Wherein, the method comprises the steps of, wherein,(i,j＝1,2,……,N)，f _i (f _i > 0) represents the i-th service-related information corresponding to the usage characteristic index, f _j (f _j > 0) represents the service-related information corresponding to the jth usage characteristic index; thenThus the pair-wise comparison matrix m= (M _ij ) _N×N And necessarily a uniform matrix.

According to the embodiment, the pair comparison matrix is constructed by the analytic hierarchy process based on the service related information, so that the objectivity of assignment of the pair comparison matrix can be improved, and the accuracy of the weight corresponding to each finally obtained use characteristic index is improved.

S204, determining respective service recommendation weights of the plurality of using characteristic indexes according to the index weight representation data.

In the embodiment of the present disclosure, the service recommendation weight of each usage characteristic index may represent the importance degree of the corresponding usage characteristic index for the service recommendation degree.

In a specific embodiment, the index weight characterizing data may include: the pair-wise comparing the matrix, determining the service recommendation weights of the plurality of usage characteristic indexes according to the index weight characterizing data may include:

1) And determining a weight vector according to the feature vector corresponding to the maximum feature value of the paired comparison matrix.

Schematically, according to user activity index F ₁ Teleconference initiation index F ₂ Shared document creation index F ₃ The obtained usage characteristic index sequence F= { F ₁ ，F ₂ ，F ₃ For example, a pair-wise comparison matrix is generated Solving the eigenvalue and eigenvector of the pair of comparison matrices, and taking the maximum eigenvalue lambda _max And its corresponding feature vector x _max ＝{x ₁ ，x ₂ ，x ₃ -apply the feature vector x _max As a weight vector.

2) And determining the service recommendation weight of each of the plurality of using characteristic indexes according to the weight vector.

In a specific embodiment, determining the service recommendation weight of each of the plurality of usage characteristic indicators according to the weight vector may include: and carrying out normalization processing on a plurality of elements in the weight vector to obtain respective business recommendation weights of the plurality of using characteristic indexes.

Illustratively, user liveness index F ₁ Is a business recommendation weight of (1)Teleconferencing initiation index F ₂ Is>Shared document creation index F ₃ Is a business recommendation weight of (1)

The above embodiment can be seen that, by determining the weight vector by the pair comparison matrix constructed by the hierarchical analysis method based on the service related information, the service recommendation weights of the plurality of service characteristic indexes are obtained, and the accuracy of the service recommendation weights can be improved.

S205, based on the service recommendation weight, the index feature information of the plurality of using feature indexes corresponding to each service object is integrated, and the service recommendation information of each service object for the resource service is obtained.

In the embodiment of the present disclosure, the service recommendation information of each service object for the resource service may characterize the service recommendation degree of the corresponding service object for the resource service. Generally, the higher the service recommendation degree of a certain service object, the greater the probability of the service object to perform subsequent service interaction with the resource service is considered.

In an alternative embodiment, the presentation form of the service recommendation information may be a service recommendation score. Schematically, a certain business object is enterprise A, and user activity index F corresponding to enterprise A ₁ Index feature information of (d) _(A,1) Teleconference initiation index F corresponding to enterprise A ₂ Index feature information of (d) _(A,2) Shared document creation index F corresponding to enterprise A ₃ Index feature information of (d) _(A,3) Business recommendation score=p for enterprise a ₁ ×d _(A,1) +p ₂ ×d _(A,2) +p ₃ ×d _(A,3) 。

S206, recommending the resource service to the plurality of service objects based on the service recommendation information of each of the plurality of service objects.

In a specific embodiment, the recommending the resource service for the plurality of service objects based on the service recommendation information of each of the plurality of service objects may include:

1) Sequencing the plurality of business objects based on the business recommendation information of each of the plurality of business objects to obtain recommendation sequence information;

2) And recommending the resource service to the plurality of service objects according to the recommendation sequence information.

Illustratively, taking service recommendation information of each of a plurality of service objects as service recommendation division as an example, sorting the plurality of service objects from large to small according to the service recommendation division to obtain recommendation sequence information, and carrying out marketing recommendation for resource service on a target number of service objects ranked at the front according to the recommendation sequence information, wherein the target number is less than or equal to the number of the plurality of service objects.

1) Taking a business object, corresponding to the business recommendation information, of the plurality of business objects as a target business object, wherein the business object meets preset business recommendation conditions;

2) And recommending the resource service to the target service object.

Specifically, the preset service recommendation condition can be preset by combining the range of the service recommendation information and the accuracy requirement of the service object screening. For example, the service recommendation information (service recommendation score) of the plurality of service objects ranges from 0 to 1, and the preset service recommendation condition may be set such that the service recommendation score is greater than 0.6, i.e., the service object corresponding to the service recommendation score greater than 0.6 may be regarded as the target service object.

According to the embodiment, based on the service recommendation information of each of the plurality of service objects, the plurality of service objects are sequenced to obtain recommendation sequence information, and the service objects with potential of service interaction are mined according to the recommendation sequence information, so that the service recommendation is performed on the potential objects in a targeted manner, and the success rate of subsequent service interaction can be effectively improved.

In some embodiments, the technical solution provided by the present application may be applied to a service recommendation scenario based on potential service object mining, referring to fig. 10, taking a service object as an enterprise and a resource service as a software product as an example, and fig. 10 is a flow chart of a service recommendation method based on potential enterprise client mining provided by the embodiment of the present application. Specifically, the method may include:

s1001, a plurality of enterprise clients with high activity are screened from an enterprise use characteristic database of the software product.

Specifically, a user usage feature database of a software product may be obtained, where the user usage feature database may include a portion of users labeled with an enterprise tag and a portion of users labeled with unknown enterprise tags, a graph network is constructed by taking an interaction relationship between users based on the software product as an edge, an enterprise to which the users labeled with the enterprise tags belong is transferred to the users labeled with the unknown enterprise tags, and enterprise tags of a portion of users labeled with the unknown enterprise are obtained by taking an enterprise name as a primary key.

Alternatively, the business names of the business using the feature database are not all standardized (e.g., there are non-standard business names that are self-populated by the customer), and thus, the non-standard business names may be converted to standardized business names by calling the API interface for the business name standardization. For example, the non-standardized business name is: tencerting, it translates to a standardized business name: shenzhen City Tencentrated computer systems Limited.

S1002, index feature information of a plurality of corresponding usage characteristic indexes of each enterprise client is generated according to the enterprise usage feature database.

Specifically, a plurality of usage characteristic indexes can be preset according to product characteristics of the software product, and index characteristic information of each of a plurality of usage characteristic indexes corresponding to each enterprise client is determined according to an enterprise usage characteristic database.

S1003, index feature information of each business interaction index of a plurality of enterprise clients is generated according to the business interaction information of the software product.

Specifically, service interaction information (for example, enterprise order information of payment software) of the software product can be obtained, enterprise clients, which have undergone service interaction with the software product, in a plurality of enterprise clients are taken as positive samples, the positive samples are marked, and index feature information of respective service interaction indexes of the plurality of enterprise clients is generated according to the service interaction information. Illustratively, if the enterprise client A does not purchase the payment version software, the index characteristic information of the service interaction index of the enterprise client A defaults to zero; after purchasing the payment version software, the enterprise client B can determine index characteristic information of the business interaction index according to the amount of the deposit.

S1004, determining respective service recommendation weights of a plurality of using characteristic indexes by using a hierarchical analysis method based on service related information, thereby obtaining service recommendation scores of each enterprise client.

Specifically, the index feature information of each of a plurality of usage characteristic indexes corresponding to each of a plurality of enterprise clients may be integrated to obtain an enterprise index feature sequence of each of the plurality of usage characteristic indexes, where each of the enterprise index feature sequences of the usage characteristic indexes may include index feature information of a plurality of enterprise clients corresponding to the usage characteristic indexes; correspondingly, index feature information of the business interaction indexes of the enterprise clients can be integrated, so that an enterprise index feature sequence of the business interaction indexes is obtained. And then, carrying out business correlation analysis on the enterprise index feature sequences of the plurality of use characteristic indexes and the enterprise index feature sequences of the business interaction indexes by using a hierarchical analysis method based on business correlation information to obtain business correlation information of the plurality of use characteristic indexes, and generating a pair comparison matrix corresponding to the plurality of use characteristic indexes according to the business correlation information, thereby determining respective business recommendation weights of the plurality of use characteristic indexes according to the feature vector corresponding to the maximum feature value of the pair comparison matrix, and carrying out fusion processing on the index feature information of the plurality of use characteristic indexes of each enterprise customer to obtain business recommendation scores of each enterprise customer.

Optionally, the algorithm of the step may be packaged to obtain a service recommendation score calculation model, and the index feature information of each of the plurality of enterprise clients is input into the service recommendation score calculation model, so that the service recommendation score of each enterprise client may be automatically input.

S1005, recommending software products to potential enterprise clients based on the service recommendation score.

Specifically, business recommendation can be divided into enterprise customers with higher business interaction potential, and the enterprise customers can be preferentially provided for sales personnel in corresponding industries or areas, and behavior characteristics of the enterprise can be provided for sales personnel after being reprocessed, so that the enterprise customers are assisted to know actual demand points of the enterprise customers for software products, and the effect of business recommendation, namely the success rate of subsequent business interaction with potential enterprise customers, is improved.

S1006, based on the business interaction behavior of the potential enterprise client, updating the business interaction information of the software product, thereby expanding the positive sample pool.

A potential business customer can be enabled to take new business interactions (e.g., generate new orders) due to a single execution run of the business recommendation. Therefore, in the next service recommendation period, a new service interaction enterprise can be added into the data input to carry out a new round of positive sample labeling, so that a positive sample pool is expanded. Meanwhile, the new business interaction behavior also supplements the demand characteristic information of richer business clients, forms positive feedback circulation, optimizes the learning of the algorithm model layer on the purchasing reasons and rules of the clients, outputs more accurate and wide potential business clients, and brings more newly-added business interaction enterprises.

Referring to table 1, the effect comparison of service recommendation performed by using the method provided by the application and the related art method is shown, and specifically, the related art includes: SVM (support vector machine), random forest algorithm, and Xgboost algorithm (extreme gradient lifting tree algorithm). It can be obviously seen that the method provided by the application has higher accuracy and recall rate in the scene.

	SVM	Random forest	Xgboost	The method of the application
					Accuracy rate of	5.0％	5.5％	5.1％	7.4％
Recall rate of recall	65.8％	63.2％	65.8％	85.8％

TABLE 1

Referring to table 2, a comparison of service interaction effects after service recommendation is performed by using the method provided by the present application and the related art method in one embodiment is shown. Specifically, the service interaction effect is shown as follows: and taking k% of service objects before the service recommendation sorting to generate the success rate of service interaction after service recommendation. It can be obviously found that the recommendation sub-ordering given by the method provided by the application in the scene has better service interaction success rate.

	SVM	Random forest	Xgboost	The method of the application
					Top 5% success rate	13.2％	2.63％	5.26％	24.1％
Top 10% success rate	23.7％	13.2％	15.8％	41.4％
					Top 25% success rate	44.7％	26.3％	26.3％	65.4％

TABLE 2

According to the technical scheme provided by the embodiment of the application, the subordination clustering processing can be carried out on the plurality of sub-objects based on the resource interaction relation, at least one sub-object corresponding to each of the plurality of service objects is determined, the use characteristic extraction is carried out on the resource use information of at least one sub-object corresponding to each of the plurality of service objects, the resource use characteristic information corresponding to each of the plurality of service objects is obtained, the subordination sub-object corresponding to each of the service objects can be clustered accurately, the accuracy of the resource use characteristic information corresponding to each of the service objects is improved, the index characteristic information of a plurality of use characteristic indexes corresponding to each of the plurality of service objects and the index characteristic information of the service interaction indexes corresponding to each of the plurality of service objects are determined according to the preset use characteristic indexes and service interaction indexes, then the correlation coefficient between each use characteristic index and the service interaction indexes or the characteristic importance characterization data corresponding to each of the use characteristic indexes is used as the service related information of each of the use characteristic indexes, the paired objective analysis matrix based on the service related information can be constructed, and the paired objective matrix can be compared and the comparison matrix can be avoided, and the comparison matrix is not interfered by the subjective and the comparison matrix is improved, and the comparison matrix is not correspondingly compared; and finally, according to the service recommendation weight determined by the pair comparison matrix, the index feature information of a plurality of service indexes corresponding to each service object is fused, so that the accuracy of the service recommendation information of the service object can be improved, the service objects with service interaction potential can be mined by combining the sequencing of the service recommendation information, the service recommendation can be performed on the potential objects in a targeted manner, and the success rate of subsequent service interaction can be effectively improved.

The embodiment of the application also provides a service recommending device, as shown in fig. 11, the service recommending device may include:

the information obtaining module 1110 is configured to obtain resource usage feature information for a resource service corresponding to each of the plurality of service objects and service interaction feature information for the resource service corresponding to each of the plurality of service objects, where the resource usage feature information includes: index feature information of each of the plurality of usage characteristic indexes, the service interaction feature information includes: index characteristic information of the business interaction index;

the service correlation analysis module 1120 is configured to perform service correlation analysis on index feature information of a plurality of service usage characteristic indexes corresponding to a plurality of service objects and index feature information of service interaction indexes corresponding to a plurality of service objects, so as to obtain service correlation information corresponding to each of the plurality of service usage characteristic indexes, where the service correlation information corresponding to each of the service usage characteristic indexes characterizes importance of the corresponding service usage characteristic index to the service interaction index;

the index weight representation data generating module 1130 is configured to generate index weight representation data based on service related information corresponding to each of the plurality of usage characteristic indexes, where each parameter in the index weight representation data is used to represent a relative importance of any two usage characteristic indexes in the plurality of usage characteristic indexes to the service interaction index;

A service recommendation weight determining module 1140, configured to determine respective service recommendation weights of the plurality of usage characteristic indicators according to the indicator weight characterization data;

the service recommendation information module 1150 is configured to perform fusion processing on index feature information of a plurality of usage characteristic indexes corresponding to each service object based on the service recommendation weight, so as to obtain service recommendation information of each service object for the resource service;

the service recommendation module 1160 is configured to recommend a resource service to a plurality of service objects based on service recommendation information of each of the plurality of service objects.

In a specific embodiment, the information obtaining module 1110 may include:

the sub-object information acquisition unit is used for acquiring resource usage information corresponding to each of the plurality of sub-objects and resource interaction relation corresponding to the resource service among the plurality of sub-objects;

the subordinate clustering processing unit is used for performing subordinate clustering processing on the plurality of sub-objects based on the resource interaction relationship to obtain a plurality of sub-object groups, at least one sub-object in each sub-object group belongs to the same target object, and the target object is any one of a plurality of service objects;

The sub-object determining unit is used for determining at least one sub-object corresponding to each of the plurality of business objects according to the plurality of sub-object groups;

and the use feature extraction unit is used for extracting the use feature of the resource use information of at least one sub-object corresponding to each of the plurality of business objects to obtain the resource use feature information corresponding to each of the plurality of business objects.

In a specific embodiment, the plurality of sub-objects may include a plurality of first sub-objects and a plurality of second sub-objects, where the plurality of first sub-objects are sub-objects with subordinate object information marked therein, and the plurality of second sub-objects are sub-objects with no subordinate object information marked therein, and the subordinate cluster processing unit may include:

the sub-object relation network diagram generating unit is used for generating a sub-object relation network diagram by taking a plurality of sub-objects as nodes and the resource interaction relation as an edge;

the label propagation unit is used for carrying out label propagation on the plurality of second sub-objects according to the sub-object relation network graph and the subobject information of the plurality of first sub-objects to obtain subobject information of the plurality of second sub-objects;

and the sub-object group unit is used for performing subordinate clustering processing on the plurality of sub-objects according to the subordinate object information of the plurality of sub-objects to obtain a plurality of sub-object groups.

In a specific embodiment, the service correlation analysis module 1120 may include:

the correlation coefficient determining unit is used for determining the correlation coefficient between the target index and the service interaction index according to the index characteristic information of the target index corresponding to the service objects and the index characteristic information of the service interaction index corresponding to the service objects, wherein the target index is any one of the service characteristic indexes;

and the first service related information unit is used for taking the corresponding correlation coefficient of each of the plurality of using characteristic indexes as service related information corresponding to each of the plurality of using characteristic indexes.

the model training unit is used for carrying out service interaction index prediction training on the tree model to be trained by taking index feature information of a plurality of service interaction indexes corresponding to each service object as a training sample and taking index feature information of the service interaction indexes corresponding to each service object as a classification result corresponding to the training sample to obtain a trained tree model;

the feature importance analysis unit is used for carrying out feature importance analysis on the plurality of using characteristic indexes based on the trained tree model to obtain feature importance representation data corresponding to the plurality of using characteristic indexes;

And the second service related information unit is used for using the characteristic importance degree representation data corresponding to each of the plurality of using characteristic indexes as service related information corresponding to each of the plurality of using characteristic indexes.

In a specific embodiment, the index weight characterization data generation module 1130 may include:

the using characteristic index sequence unit is used for obtaining a using characteristic index sequence according to a plurality of using characteristic indexes;

the index acquisition unit is used for acquiring any one first index and any one second index, wherein the first index and the second index are the use characteristic indexes in the use characteristic index sequence;

the target parameter determining unit is used for determining a target parameter according to the service related information corresponding to the first index and the service related information corresponding to the second index, wherein the target parameter represents the relative importance of the first index to the service interaction index compared with the second index;

a parameter position determining unit, configured to determine a position of the target parameter in the index weight characterization data according to a position of the first index in the usage characteristic index sequence and a position of the second index in the usage characteristic index sequence;

the data generation unit is used for generating index weight representation data according to the target parameters and the positions of the target parameters in the index weight representation data.

In a specific embodiment, the index weight characterizing data may include: the pair-wise comparison matrix, the service recommendation weight determining module 1140 may include:

the weight vector determining unit is used for determining a weight vector according to the feature vector corresponding to the maximum feature value of the paired comparison matrixes;

and the weight determining unit is used for determining the service recommendation weight of each of the plurality of using characteristic indexes according to the weight vector.

In a specific embodiment, the service recommendation module 1160 may include:

a recommendation sequence information unit, configured to sort the plurality of service objects based on respective service recommendation information of the plurality of service objects, to obtain recommendation sequence information;

and the recommending unit is used for recommending the resource service for the plurality of service objects according to the recommending sequence information.

It should be noted that the apparatus and method embodiments in the apparatus embodiments are based on the same inventive concept.

The embodiment of the application provides a service recommending device, which comprises a processor and a memory, wherein at least one instruction or at least one section of program is stored in the memory, and the at least one instruction or the at least one section of program is loaded and executed by the processor to realize the service recommending method provided by the embodiment of the method.

Further, fig. 12 shows a schematic hardware structure of a service recommendation device for implementing the service recommendation method provided by the embodiment of the present application, where the service recommendation device may participate in forming or including the service recommendation apparatus provided by the embodiment of the present application. As shown in fig. 12, the service recommendation device 120 may include one or more (shown with 1202a, 1202b, … …,1202 n) processors 1202 (the processor 1202 may include, but is not limited to, a microprocessor MCU or a processing device such as a programmable logic device FPGA), a memory 1204 for storing data, and a transmission 1206 for communication functions. In addition, the method may further include: a display, an input/output interface (I/O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of the I/O interface), a network interface, a power supply, and/or a camera. It will be appreciated by those of ordinary skill in the art that the configuration shown in fig. 12 is merely illustrative and is not intended to limit the configuration of the electronic device described above. For example, the service recommendation device 120 may also include more or fewer components than shown in fig. 12, or have a different configuration than shown in fig. 12.

It should be noted that the one or more processors 1202 and/or other data processing circuits described above may be referred to herein generally as "data processing circuits. The data processing circuit may be embodied in whole or in part in software, hardware, firmware, or any other combination. Further, the data processing circuitry may be a single stand-alone processing module, or incorporated in whole or in part into any of the other elements in the business recommendation device 120 (or mobile device). As referred to in embodiments of the application, the data processing circuit acts as a processor control (e.g., selection of the path of the variable resistor termination connected to the interface).

The memory 1204 may be used for storing software programs and modules of application software, and the processor 1202 executes the software programs and modules stored in the memory 1204 to perform various functional applications and data processing, i.e., to implement a service recommendation method according to the embodiments of the present application. Memory 1204 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 1204 may further include memory remotely located with respect to the processor 1202, which may be connected to the service recommendation device 120 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmission means 1206 is used to receive or transmit data via a network. The specific example of the network described above may include a wireless network provided by a communication provider of the service recommendation device 120. In one example, the transmission means 1206 includes a network adapter (NetworkInterfaceController, NIC) that can be connected to other network devices via a base station to communicate with the internet. In one embodiment, the transmission means 1206 may be a radio frequency (RadioFrequency, RF) module for communicating wirelessly with the internet.

The display may be, for example, a touch screen type Liquid Crystal Display (LCD) that may enable a user to interact with a user interface of the service recommendation device 120 (or mobile device).

Embodiments of the present application also provide a computer readable storage medium, which may be provided in a service recommendation device to store at least one instruction or at least one program related to a service recommendation method in a method embodiment, where the at least one instruction or the at least one program is loaded and executed by the processor to implement the service recommendation method provided in the method embodiment.

Alternatively, in this embodiment, the storage medium may be located in at least one network server among a plurality of network servers of the computer network. Alternatively, in the present embodiment, the storage medium may include, but is not limited to: a U-disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing program codes.

Embodiments of the present application also provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions to cause the computer device to perform the business recommendation method as provided by the method embodiments.

It should be noted that: the sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this application. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

The embodiments of the present application are described in a progressive manner, and the same and similar parts of the embodiments are all referred to each other, and each embodiment is mainly described in the differences from the other embodiments. In particular, for the apparatus and device embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, with reference to the description of the method embodiments in part.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program for instructing relevant hardware, where the program may be stored in a computer readable storage medium, and the storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The foregoing description of the preferred embodiments of the application is not intended to limit the application to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the application are intended to be included within the scope of the application.

Claims

1. A business recommendation method, the method comprising:

2. The method of claim 1, wherein performing service correlation analysis on the index feature information of the service interaction indexes corresponding to the service objects and the index feature information of the service interaction indexes corresponding to the service objects to obtain service correlation information corresponding to the service interaction indexes respectively comprises:

determining a correlation coefficient between the target index and the service interaction index according to index characteristic information of the target index corresponding to the service objects and index characteristic information of the service interaction index corresponding to the service objects, wherein the target index is any one of the service characteristic indexes;

and taking the correlation coefficient corresponding to each of the plurality of using characteristic indexes as service correlation information corresponding to each of the plurality of using characteristic indexes.

3. The method of claim 1, wherein performing service correlation analysis on the index feature information of the service interaction indexes corresponding to the service objects and the index feature information of the service interaction indexes corresponding to the service objects to obtain service correlation information corresponding to the service interaction indexes respectively comprises:

Taking index feature information of a plurality of service interaction indexes corresponding to each service object as a training sample, taking the index feature information of the service interaction indexes corresponding to each service object as a classification result corresponding to the training sample, and carrying out service interaction index prediction training on a tree model to be trained to obtain a trained tree model;

based on the trained tree model, carrying out feature importance analysis on the plurality of using characteristic indexes to obtain feature importance characterization data corresponding to the plurality of using characteristic indexes;

and taking the characteristic importance degree representation data corresponding to each of the plurality of using characteristic indexes as service related information corresponding to each of the plurality of using characteristic indexes.

4. The method of claim 1, wherein generating the metric weight characterization data based on the traffic-related information corresponding to each of the plurality of usage characteristic metrics comprises:

obtaining a use characteristic index sequence according to the plurality of use characteristic indexes;

acquiring any one first index and any one second index, wherein the first index and the second index are the use characteristic indexes in the use characteristic index sequence;

Determining a target parameter according to the service related information corresponding to the first index and the service related information corresponding to the second index, wherein the target parameter represents the relative importance of the first index to the service interaction index compared with the second index;

determining the position of the target parameter in the index weight representation data according to the position of the first index in the using characteristic index sequence and the position of the second index in the using characteristic index sequence;

and generating the index weight representation data according to the target parameter and the position of the target parameter in the index weight representation data.

5. The method of claim 1, wherein the metric weight characterization data comprises: and a pair-wise comparison matrix, wherein the determining the service recommendation weights of the plurality of usage characteristic indexes according to the index weight characterization data comprises:

determining a weight vector according to the feature vector corresponding to the maximum feature value of the paired comparison matrix;

and determining the service recommendation weight of each of the plurality of use characteristic indexes according to the weight vector.

6. The method of claim 1, wherein the obtaining the resource usage characteristic information for the resource service corresponding to each of the plurality of service objects includes:

Acquiring resource usage information corresponding to each of a plurality of sub-objects and aiming at the resource service and resource interaction relation among the plurality of sub-objects;

performing subordinate clustering processing on the plurality of sub-objects based on the resource interaction relationship to obtain a plurality of sub-object groups, wherein at least one sub-object in each sub-object group is subordinate to the same target object, and the target object is any service object in the plurality of service objects;

determining at least one sub-object corresponding to each of the plurality of business objects according to the plurality of sub-object groups;

extracting the use characteristics of the resource use information of at least one sub-object corresponding to each of the plurality of business objects to obtain the resource use characteristic information corresponding to each of the plurality of business objects.

7. The method of claim 6, wherein the plurality of sub-objects includes a plurality of first sub-objects and a plurality of second sub-objects, the plurality of first sub-objects are sub-objects labeled with subordinate object information in the plurality of sub-objects, the plurality of second sub-objects are sub-objects not labeled with subordinate object information in the plurality of sub-objects, the performing subordinate clustering on the plurality of sub-objects based on the resource interaction relationship, and the obtaining a plurality of sub-object groups includes:

Generating a sub-object relationship network diagram by taking the plurality of sub-objects as nodes and the resource interaction relationship as an edge;

according to the sub-object relation network diagram and the subobject information of the plurality of first sub-objects, carrying out tag propagation on the plurality of second sub-objects to obtain subobject information of the plurality of second sub-objects;

and performing subordinate clustering processing on the plurality of sub-objects according to the subordinate object information of the plurality of sub-objects to obtain the plurality of sub-object groups.

8. The method according to any one of claims 1 to 7, wherein the recommending the plurality of business objects for the resource business based on the respective business recommendation information of the plurality of business objects includes:

sequencing the plurality of business objects based on the business recommendation information of each of the plurality of business objects to obtain recommendation sequence information;

and recommending the resource service to the plurality of service objects according to the recommendation sequence information.

9. A service recommendation device, characterized in that the device comprises:

10. A service recommendation device, characterized in that the device comprises a processor and a memory, in which at least one instruction or at least one program is stored, which is loaded and executed by the processor to implement the service recommendation method according to any of claims 1 to 8.

11. A computer readable storage medium having stored therein at least one instruction or at least one program, the at least one instruction or the at least one program being loaded and executed by a processor to implement the service recommendation method of any one of claims 1 to 8.

12. A computer program product, characterized in that it comprises at least one instruction or at least one program, which is loaded and executed by a processor to implement the service recommendation method according to any of claims 1 to 8.