CN114202418A - Information processing method, device, equipment and medium - Google Patents
Information processing method, device, equipment and medium Download PDFInfo
- Publication number
- CN114202418A CN114202418A CN202111502592.3A CN202111502592A CN114202418A CN 114202418 A CN114202418 A CN 114202418A CN 202111502592 A CN202111502592 A CN 202111502592A CN 114202418 A CN114202418 A CN 114202418A
- Authority
- CN
- China
- Prior art keywords
- transaction
- graph
- social entity
- social
- entity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/12—Accounting
- G06Q40/123—Tax preparation or submission
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Marketing (AREA)
- Biomedical Technology (AREA)
- General Business, Economics & Management (AREA)
- Technology Law (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Strategic Management (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present disclosure relates to an information processing method, apparatus, device, and medium. An information processing method comprising: acquiring a transaction graph of a social entity set, wherein the transaction graph comprises basic information of all social entities in the social entity set and transaction relations among all the social entities; screening and acquiring a transaction sub-graph containing a target social entity from the transaction graph; and determining a target social entity to be recommended from the transaction subgraph through a link prediction model. Therefore, the data volume can be effectively reduced and the calculation speed is accelerated for various social entity recommendation algorithms through the transaction sub-graph generation method, and the deep mining of the social entity transaction network can be realized by recommending the social entities through the link prediction model, so that the accuracy of the recommendation algorithm is improved.
Description
Technical Field
The present disclosure relates to the field of information processing technologies, and in particular, to an information processing method, apparatus, device, and medium.
Background
With the development of the tax big data, the related departments accumulate massive social entity transaction data, and the potential value in the tax big data is efficiently mined to become the direction of important attention of users of the related social entities. For the social entity, the commercial profit of the social entity is closely related to the customer information, and the selection of the upstream and downstream customers can ensure the stable profit and the reduced risk of the social entity. Therefore, accurate recommendation of high-quality customers for social entities is a crucial link in social entity operation.
The currently disclosed social entity recommendation methods include a recommendation method based on content matching, and the method uses characteristics such as social entity industry characteristics, social entity geographical characteristics, social entity transaction information and the like to perform recommendation in a similarity calculation manner. The recommendation method has strong interpretability of the recommendation result and stable recommendation result, but cannot mine deep recommendation relation among social entities, and in addition, the problem that the calculation amount is large and effective calculation cannot be realized due to the fact that tax massive data exist is solved.
Disclosure of Invention
To solve the technical problems described above or at least partially solve the technical problems, the present disclosure provides an information processing method, apparatus, device, and medium.
In a first aspect, an embodiment of the present disclosure provides an information processing method, including:
acquiring a transaction graph of a social entity set, wherein the transaction graph comprises basic information of all social entities in the social entity set and transaction relations among all the social entities;
screening and acquiring a transaction sub-graph containing a target social entity from the transaction graph;
and determining a target social entity to be recommended from the transaction subgraph through a link prediction model.
In a second aspect, an embodiment of the present disclosure provides an information processing apparatus, including:
the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring a transaction graph of a social entity set, and the transaction graph comprises basic information of all social entities in the social entity set and transaction relations among all the social entities;
the extraction unit is used for screening and acquiring a transaction sub-graph containing the target social entity from the transaction graph;
and the determining unit is used for determining the target social entity to be recommended from the transaction subgraph through the link prediction model.
In a third aspect, an embodiment of the present disclosure provides an information processing apparatus, including:
a processor;
a memory for storing executable instructions;
the processor is configured to read the executable instructions from the memory and execute the executable instructions to implement the information processing method according to the first aspect.
In a fourth aspect, the present disclosure provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the computer program causes the processor to implement the information processing method according to the first aspect.
Compared with the prior art, the technical scheme provided by the embodiment of the disclosure has the following advantages:
according to the information processing method, the device, the equipment and the medium, after the transaction graph of the social entity set is obtained, the transaction subgraph containing the target social entity is screened out from the transaction graph, and the target social entity to be recommended is obtained from the transaction subgraph through the link prediction model, so that the data volume can be effectively reduced for various social entity recommendation algorithms through the transaction subgraph generation method, the calculation speed is accelerated, deep mining of the social entity transaction network can be realized through social entity recommendation through the link prediction model, and the accuracy of the recommendation algorithm is improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present disclosure, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
Fig. 1 is a schematic flowchart of an information processing method according to an embodiment of the present disclosure;
fig. 2 is a schematic structural diagram of an information processing apparatus according to an embodiment of the present disclosure;
fig. 3 is a schematic diagram of a hardware circuit structure of an information processing apparatus according to an embodiment of the present disclosure.
Detailed Description
In order that the above objects, features and advantages of the present disclosure may be more clearly understood, aspects of the present disclosure will be further described below. It should be noted that the embodiments and features of the embodiments of the present disclosure may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure, but the present disclosure may be practiced in other ways than those described herein; it is to be understood that the embodiments disclosed in the specification are only a few embodiments of the present disclosure, and not all embodiments.
With the development of tax big data, related departments accumulate massive social entity transaction data, the potential value in the tax big data is efficiently mined to become the direction of important attention of users of related social entities, and the currently disclosed social entity recommendation method comprises a recommendation method based on content matching and a method based on link prediction.
The method based on the link prediction belongs to a new recommendation method, and can automatically mine deep relationships among social entities. However, for massive social entity data, the link prediction method faces the problems of too large calculation amount and incapability of implementation. In addition, the method is difficult to train negative sample data construction.
These detection schemes suffer from the following disadvantages:
firstly, the recommendation system caused by the mass data in the tax field has low calculation efficiency.
Secondly, the deep learning recommendation system lacks negative sample data recommended by social entities.
Finally, the sample construction scheme based on traditional machine learning cannot filter noisy data in training data, so that the model trained based on high-noise training data has poor detection accuracy, and is difficult to accurately recommend social entities.
In conclusion, the existing method for accurately recommending social entities is difficult to meet the actual business requirements of tax social entities for screening high-quality customers from mass data.
In order to solve the above problems, embodiments of the present disclosure provide an information processing method, apparatus, device, and medium, which can screen a transaction graph including a target social entity from a transaction graph after obtaining the transaction graph of a social entity set, and obtain the target social entity to be recommended from the transaction graph through a link prediction model, so that data volume can be effectively reduced for various social entity recommendation algorithms through a transaction graph generation method, computation speed is accelerated, deep mining of a social entity transaction network can be realized through social entity recommendation through the link prediction model, and accuracy of a recommendation algorithm is improved.
First, an information processing method provided by an embodiment of the present disclosure is described below with reference to fig. 1 to 3.
Fig. 1 shows a schematic flowchart of an information processing method provided by an embodiment of the present disclosure.
In some embodiments of the present disclosure, the information processing method illustrated in fig. 1 may be performed by a server. The server may be a cloud server or a server cluster or other devices with storage and computing functions.
As shown in fig. 1, the information processing method may include the following steps.
And S110, acquiring a transaction graph of the social entity set.
In the embodiment of the disclosure, the server may obtain a transaction graph of the social entities, where the transaction graph includes basic information of each social entity in the social entity set and transaction relationships between each social entity.
The generation process of the trading map is as follows:
in the production and operation activities of the social entities, due to the trading behaviors of selling goods, purchasing commodities and the like, the social entities form a complex trading relationship.
In the specific construction process of the transaction graph, the social entities can be used as nodes, and the transaction relationships among the social entities can be used as edges, so that the transaction relationships among the social entities can be represented as a huge network graph structure, namely the transaction graph. The nodes in the graph carry basic information of the social entities and carry transaction information among the social entities.
The social entity may include a tax-paying social entity, and the basic information of the social entity may include registration data, industry, business address, credit rating, and the like of the social entity.
The transaction relationship between the social entities may include information such as transaction amount, transaction time, transaction item, etc.
Alternatively, the registered registration data may include at least data of a registered address of the target social entity, corporate information, an entity name, a declaration business scope, financial staff information, stockholder information, part-time staff information, and the like.
And S120, screening and acquiring a transaction sub-graph containing the target social entity from the transaction graph.
In the embodiment of the disclosure, after obtaining the transaction graph of the social entity, the server extracts, according to a certain screening condition, a transaction sub-graph containing the target social entity, that is, a part of graph structure of the transaction graph, from the transaction graph, and the transaction sub-graph can be used for quickly determining the target social entity.
The number of nodes in the social entity transaction graph is large, and if the recommended target social entity is directly calculated on the whole graph structure, the problems of extremely high calculation complexity and extremely large calculation resource occupation can be met. In addition, since a vast majority of the transaction relationships between social entities occur in the same industry or in upstream and downstream industry chains, computing on all social entity nodes results in a large amount of useless computing. Therefore, the calculation complexity is reduced by constructing a transaction subgraph, and the algorithm usability is improved.
In this embodiment of the present disclosure, optionally, the screened transaction subgraph may specifically include: at least one of an industry transaction sub-graph, a downstream industry transaction sub-graph and a social entity local transaction sub-graph;
the industry transaction subgraph is constructed by screening industry information of social entities; for example: and constructing the 'hardware retail' industry subgraph, traversing all graph nodes, deleting the social entity nodes and connecting edges thereof which are not 'hardware retail' in the industry, and obtaining the constructed industry transaction subgraph by the remaining nodes and the connecting edges.
The business transaction subgraph is formed by screening and constructing social entities through pre-constructed business information of the upstream and downstream; for example: and constructing an upstream and downstream industry subgraph of hardware processing and selling, traversing all graph nodes, deleting social entity nodes and connecting edges thereof in the node industry which are not in the upstream and downstream industry lists, and taking the rest nodes and connecting edges as the constructed subgraph.
The social entity local transaction subgraph is constructed by carrying out a graph search algorithm on a target social entity, and is the subgraph with the finest granularity. For example, a transaction sub-graph of the social entity a is constructed, and first, for the node social entity a, a social entity node with a depth of n is searched in the transaction graph by using a depth-first search algorithm or a breadth-first search algorithm. And then, utilizing the searched social entity node set to retrieve a transaction subgraph only containing edges connecting the set nodes and the set nodes from the transaction graph.
Further, when the transaction sub-graph is a social entity local transaction sub-graph, the transaction graph of the social entity further includes: and the information of the commodities transacted with each social entity is collected with the social entities.
The transaction subgraph constructed by the method of the social entity local transaction subgraph has the problem of social entity omission. Assuming that both social entity a and social entity B have transaction requirements for commodity C, the route from social entity a to social entity B is equal to 3. When the search depth is 2, the social entity B is excluded from the local transaction subgraph of the social entity a, so that information is lost, and finally, a recommendation result is lost.
Therefore, in the embodiment of the disclosure, the social entity and commodity node difference graph is used for optimizing the local transaction subgraph structure of the social entity. The social entity and commodity node abnormal graph is formed by adding commodity information transacted with the social entity on the basis of a social entity transaction graph, namely: and adding commodity transaction relationship edges corresponding to the commodity nodes and the social entities to show that the social entities have transacted certain commodities.
Specifically, when constructing the local subgraph of the social entity a, in addition to searching for social entity nodes by using a graph traversal algorithm, the adjacent nodes of the adjacent commodity nodes of the social entity a are also added into the social entity node set. The constructed social entity local transaction sub-graph not only can reduce the scale of the transaction graph, but also avoids the problem of social entity node omission caused by overlong historical transaction paths.
Further, when the transaction sub-graph is a social entity local transaction sub-graph, the method further comprises the following steps:
and randomly selecting a social entity set without a transaction relation with the target social entity from the transaction graph as negative sample data by a local random negative sampling method.
The link prediction model is a deep learning model, and the deep learning model needs to be trained by using positive example samples and negative example samples. However, only positive sample data representing the historical transaction of the social entity can be obtained from the historical transaction information of the social entity, and a negative sample needs to be constructed.
The random negative sampling method is to randomly select social entity nodes which are not connected with the target social entity from the global transaction graph as negative example samples. The random negative sampling is simple to implement and is widely applied to deep learning negative sample construction. However, the social entity transaction network is a sparse network, edges do not exist among most nodes, and the average path among the nodes is large. Negative examples constructed using random negative sampling have significant differences from positive examples, namely: negative example inter-node path length or no path between the candidate node and the target node. Such differences are too significant to allow the model to learn efficiently (the model may be determined by determining whether there is a short path between nodes). Aiming at the problem, the scheme provides local random sampling and negative sample construction.
The local random negative sampling method limits the sampling range on the social entity local transaction subgraph, and further, the proportion of negative samples accounts for seven tenths, namely, the sampling result is divided into the negative samples and the positive samples according to the proportion of 7: 3.
The method has two advantages:
firstly, the sampling range is limited on the social entity local subgraph, so that the positive sample and the negative sample can be prevented from having obvious path characteristic difference. Compared with the global feature subgraph, the social entity local subgraph is dense, and a short path exists between the sampled node and the target node.
Secondly, social entities in the local subgraph have similar transaction requirements, and negative sampling results have certain 'false negatives'. And marking a small part of sampling results as positive samples, and introducing noise data, thereby being beneficial to improving the generalization capability of the model.
And S130, determining a target social entity to be recommended from the transaction subgraph through a link prediction model.
In the embodiment of the disclosure, after the transaction sub-graph is acquired, the server may extract the feature information of the target social entity from the transaction sub-graph through the link prediction model, compare the feature information with the feature information of the candidate social entity, and determine whether the target social entity is to be recommended from the comparison result.
The link prediction model is used for predicting the possibility that an edge exists between two nodes without an edge connection relation in a network through the node information and the node connection relation in the network graph structure. After representing the social entities and the transaction relations among the social entities as graph structures, the social entity recommendation can be realized by a link prediction method: and predicting connecting edges which may exist in the social entity transaction graph. The link prediction model includes positive and negative example samples that need to be constructed.
The construction method provided by the embodiment of the disclosure is as follows:
given a target recommendation node social entity A, the construction process of the sample data of a positive example is as follows:
1. acquiring a social entity node set X having a connection edge relationship with a social entity A;
2. and if the node set X is not empty, taking out a node from the set X, marking as the node X, and respectively carrying out depth-first search with the depth of 2 on the node A and the node X to obtain a search subgraph G.
3. And deleting the edges of the node x and the node A in the search subgraph G. The search subgraph is added to the positive sample.
4. And repeating the steps 2) and 3) until the node assembly is empty.
Given a target recommendation node social entity A, the negative sample data construction mode is as follows:
1. sampling a group of social entity nodes X by using a local random sampling mode;
2. and if the node set X is not empty, taking out a node from the set X, marking as the node X, and respectively carrying out depth-first search with the depth of 2 on the node A and the node X to obtain a search subgraph G.
3. And calling a random number generation algorithm to generate random numbers in the range of [1-10 ]. And if the random number is larger than 3, adding the search subgraph into the negative sample, otherwise, adding the search subgraph into the positive sample.
4. And repeating the steps 2) and 3) until the node assembly is empty.
In the embodiment of the present disclosure, optionally, S130 includes: and acquiring the characteristics of the target social entity from the transaction graph by adopting a graph convolution neural network.
Further, S130 further includes determining a target social entity to recommend through the fully-connected neural network two-class.
The scheme adopts a GCN graph convolution neural network to extract the characteristics of the transaction graph. The GCN is a general graph structure feature extractor, the input of which is a graph adjacency matrix and a graph node original feature matrix, and the output of which is an extracted graph node feature matrix. And after the GCN is used for extracting the graph node feature matrix, splicing the feature vectors of the target social entity node and the candidate social entity node, and then performing secondary classification by using a fully-connected neural network. And the classification result represents whether the target social entity node and the candidate social entity node form a recommendation relation.
In the embodiment of the disclosure, the commodity name of the social entity sale item, the social entity industry code, the geographic position of the social entity, the credit rating of the social entity and the like are used as the original characteristics of the graph node.
Specifically, the commodity name of the social entity selling item can be converted into a word vector by using word2vec, the social entity industry code can be converted into a one-hot representation form, the geographic position of the social entity can be represented by using longitude and latitude, and the credit rating (excellent, good, normal, risk and abnormal) of the social entity can be represented as five ratings of 1-3. The graph adjacency matrix may be transformed from the transaction graph.
In summary, the information processing method provided by the above disclosed embodiment has the following beneficial effects:
firstly, representing the social entity transaction information as a graph data structure form of a transaction graph, providing a generation method of a social entity transaction sub-graph, and carrying out social entity prediction recommendation by using a link prediction model.
Secondly, the transaction sub-graph generation method provided by the embodiment of the disclosure can effectively reduce the data volume and accelerate the calculation speed for various social entity recommendation algorithms.
And thirdly, aiming at the problem that the training negative sample of the neural network is lacked, the method can effectively sample the training negative sample by a local negative sampling method, and avoids overfitting of model training.
Finally, the deep mining of the social entity transaction network can be realized by using the method for recommending the social entity by using the graph convolutional neural network.
Fig. 2 shows a schematic structural diagram of an information processing apparatus provided in an embodiment of the present disclosure.
In some embodiments of the present disclosure, the information processing apparatus shown in fig. 2 may be applied to a server. The server may be a cloud server or a server cluster or other devices with storage and computing functions.
As shown in fig. 2, the information processing apparatus 200 may include an acquisition unit 210, an extraction unit 220, and a determination unit 230.
The obtaining unit 210 may be configured to obtain a transaction graph of the social entity set, where the transaction graph includes basic information of each social entity in the social entity set and transaction relationships between each social entity; .
The extraction unit 220 may be configured to filter and obtain a transaction sub-graph containing the target social entity from the transaction sub-graph.
The determining unit 230 may be configured to determine the target social entity to recommend from the transaction sub-graph by means of a link prediction model.
In the embodiment of the disclosure, after the transaction graph of the social entity set is obtained, the transaction sub-graph containing the target social entity is screened out from the transaction graph, and the target social entity to be recommended is obtained from the transaction sub-graph through the link prediction model, so that the data volume can be effectively reduced and the calculation speed is accelerated for various social entity recommendation algorithms through the transaction sub-graph generation method, and the deep mining of the social entity transaction network can be realized by recommending the social entity through the link prediction model, and the accuracy of the recommendation algorithm is improved.
It should be noted that the information processing apparatus 200 shown in fig. 2 may perform each step in the method embodiment shown in fig. 1, and implement each process and effect in the method embodiment shown in fig. 1, which are not described herein again.
Fig. 3 shows a schematic diagram of a hardware circuit structure of an information processing apparatus according to an embodiment of the present disclosure.
As shown in fig. 3, the information processing apparatus 300 may include a controller 301 and a memory 302 storing computer program instructions.
Specifically, the controller 301 may include a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or may be configured to implement one or more Integrated circuits of the embodiments of the present Application.
The controller 301 reads and executes the computer program instructions stored in the memory 302 to perform the steps of the information processing method provided by the embodiment of the present disclosure.
In one example, the information processing apparatus 300 may further include a transceiver 303 and a bus 304. As shown in fig. 3, the controller 301, the memory 302 and the transceiver 303 are connected via a bus 304 to complete communication with each other.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the term "comprises/comprising" is intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
The foregoing are merely exemplary embodiments of the present disclosure, which enable those skilled in the art to understand or practice the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the disclosure. Thus, the present disclosure is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. An information processing method characterized by comprising:
acquiring a transaction graph of a social entity set, wherein the transaction graph comprises basic information of all social entities in the social entity set and transaction relations among all the social entities;
screening and acquiring a transaction sub-graph containing a target social entity from the transaction graph;
determining a target social entity to recommend from the transaction sub-graph through a link prediction model.
2. The method of claim 1, wherein the transaction sub-graph comprises:
at least one of an industry transaction sub-graph, a downstream industry transaction sub-graph and a social entity local transaction sub-graph;
the industry transaction subgraph is constructed by screening industry information of social entities;
the upstream and downstream business transaction subgraph is formed by screening and constructing social entities through pre-constructed upstream and downstream business information;
the social entity local transaction subgraph is constructed by carrying out graph search algorithm on a target social entity.
3. The method of claim 2, wherein when the transaction sub-graph is a social entity local transaction sub-graph, the transaction graph of the social entity further comprises:
and the information of the commodities transacted with each social entity is collected with the social entities.
4. The method of claim 2 or 3, wherein the transaction sub-graph is a social entity local transaction sub-graph, further comprising:
and randomly selecting a social entity set without a transaction relation with the target social entity from the transaction graph as negative sample data by a local random negative sampling method.
5. The method of claim 4, wherein the negative examples sample proportion is seven tenths.
6. The method of claim 1, wherein determining the target social entity to recommend from the transaction sub-graph through a link prediction model comprises:
and acquiring the characteristics of the target social entity from the transaction graph by adopting a graph convolution neural network.
7. The method of claim 6, wherein after obtaining the characteristics of the target social entity from the trading graph using the graph convolutional neural network, further comprising:
and determining the target social entity to be recommended through the fully-connected neural network two-class.
8. An information processing apparatus characterized by comprising:
the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring a transaction graph of a social entity set, and the transaction graph comprises basic information of all social entities in the social entity set and transaction relations among all the social entities;
the extraction unit is used for screening and acquiring a transaction sub-graph containing a target social entity from the transaction graph;
and the determining unit is used for determining the target social entity to be recommended from the transaction subgraph through a link prediction model.
9. An information processing apparatus characterized by comprising:
a processor;
a memory for storing executable instructions;
wherein the processor is used for reading the executable instructions from the memory and executing the executable instructions to realize the information processing method of any one of the above claims 1-7.
10. A computer-readable storage medium, characterized in that the storage medium stores a computer program which, when executed by a processor, causes the processor to implement the information processing method of any one of the above claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111502592.3A CN114202418A (en) | 2021-12-09 | 2021-12-09 | Information processing method, device, equipment and medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111502592.3A CN114202418A (en) | 2021-12-09 | 2021-12-09 | Information processing method, device, equipment and medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114202418A true CN114202418A (en) | 2022-03-18 |
Family
ID=80651813
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111502592.3A Pending CN114202418A (en) | 2021-12-09 | 2021-12-09 | Information processing method, device, equipment and medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114202418A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116308805A (en) * | 2023-05-25 | 2023-06-23 | 北京芯盾时代科技有限公司 | Transaction account identification method and device and electronic equipment |
-
2021
- 2021-12-09 CN CN202111502592.3A patent/CN114202418A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116308805A (en) * | 2023-05-25 | 2023-06-23 | 北京芯盾时代科技有限公司 | Transaction account identification method and device and electronic equipment |
CN116308805B (en) * | 2023-05-25 | 2023-08-08 | 北京芯盾时代科技有限公司 | Transaction account identification method and device and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110009174B (en) | Risk recognition model training method and device and server | |
US20210073283A1 (en) | Machine learning and prediction using graph communities | |
CN111291816B (en) | Method and device for carrying out feature processing aiming at user classification model | |
CN107657267B (en) | Product potential user mining method and device | |
CN111507543B (en) | Model training method and device for predicting business relation between entities | |
CN111368147B (en) | Graph feature processing method and device | |
CN111932130B (en) | Service type identification method and device | |
CN111222976A (en) | Risk prediction method and device based on network diagram data of two parties and electronic equipment | |
CN112269805A (en) | Data processing method, device, equipment and medium | |
CN114118816A (en) | Risk assessment method, device and equipment and computer storage medium | |
CN110992194A (en) | User reference index algorithm based on attribute-containing multi-process sampling graph representation learning model | |
CN114202418A (en) | Information processing method, device, equipment and medium | |
CN111245815B (en) | Data processing method and device, storage medium and electronic equipment | |
CN111476371B (en) | Method and device for evaluating specific risk faced by server | |
CN111582722B (en) | Risk identification method and device, electronic equipment and readable storage medium | |
CN112801784A (en) | Bit currency address mining method and device for digital currency exchange | |
CN110807159B (en) | Data marking method and device, storage medium and electronic equipment | |
CN111460044B (en) | Geographic position data processing method and device | |
CN110008282A (en) | Transaction data synchronization interconnection method, device, computer equipment and storage medium | |
CN112784015B (en) | Information identification method and device, apparatus, medium, and program | |
CN110570301B (en) | Risk identification method, device, equipment and medium | |
KR102282328B1 (en) | System and Method for Predicting Preference National Using Long Term Short Term Memory | |
CN113052604A (en) | Object detection method, device, equipment and storage medium | |
Fitrianto et al. | Development of direct marketing strategy for banking industry: the use of a Chi-squared Automatic Interaction Detector (CHAID) in deposit subscription classification. | |
Henriksson et al. | Housing Price Prediction over Countrywide Data: A comparison of XGBoost and Random Forest regressor models |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |