CN111581516B - Investment product recommending method and related device - Google Patents
Investment product recommending method and related device Download PDFInfo
- Publication number
- CN111581516B CN111581516B CN202010392822.4A CN202010392822A CN111581516B CN 111581516 B CN111581516 B CN 111581516B CN 202010392822 A CN202010392822 A CN 202010392822A CN 111581516 B CN111581516 B CN 111581516B
- Authority
- CN
- China
- Prior art keywords
- user
- investment
- node
- information
- investment product
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 239000013598 vector Substances 0.000 claims abstract description 98
- 238000005295 random walk Methods 0.000 claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000012544 monitoring process Methods 0.000 claims description 7
- 230000008901 benefit Effects 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 description 6
- 230000009471 action Effects 0.000 description 3
- 238000012886 linear function Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003442 weekly effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
- G06Q40/06—Asset management; Financial planning or analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Evolutionary Biology (AREA)
- Economics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Game Theory and Decision Science (AREA)
- Human Resources & Organizations (AREA)
- Operations Research (AREA)
- Evolutionary Computation (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Animal Behavior & Ethology (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application provides a recommendation method and a related device of investment products, wherein the method comprises the following steps: acquiring user information of each user and product information of each investment product; constructing a knowledge graph between the user and the investment product based on the user information of each user and the product information of each investment product; performing random walk on the knowledge graph based on a preset personal basic information priority strategy, a preset financial information priority strategy and a preset investment information priority strategy respectively to obtain a plurality of node sequences; training each node sequence by applying a word vector model to obtain the vector of the node in the node sequence; for each node sequence, calculating to obtain a relevance score between a user corresponding to each user node and an investment product corresponding to each investment product node by using a vector of the user node and a vector of the investment product node in the node sequence; and generating a recommendation list of each user according to the relevance scores between each user and each investment product.
Description
Technical Field
The application relates to the technical field of product recommendation, in particular to a recommendation method and a related device for investment products.
Background
The traditional investment consultants mainly provide investment consultation services for users in a mode of manual one-to-one communication, so that the charged rate is expensive, the investment threshold is high, and the investment consultation services are mainly used for high-net-value people. However, with the development of economy, urban and rural residents have increased in income, and the proportion of available investment financial resources in the incomes of individuals is increased, so that huge market space exists in the population of middle-aged and lower-aged products nowadays.
Because the traditional mode is no longer suitable for crowds with huge quantity, a plurality of companies now start to carry out intelligent consultation software, and data such as user behaviors, markets, products and the like are analyzed through the intelligent consultation software, so that investment combination suggestions which relatively meet the user demands are provided for the users efficiently, conveniently and at low price.
However, in the present intelligent supervision software, the weight value of each investment influencing factor is set manually, and then the information of the user and the information of the investment product are multiplied by the corresponding weight value, and the investment product with equal high score is recommended to the user. However, the situations of different users are quite various, the main factors influencing the users to purchase different investment products are different, the investment products are determined by adopting the manually set fixed weight values, the method is only applicable to the common situations, and the investment products of the most suitable users cannot be provided for the different users well.
Disclosure of Invention
Based on the shortcomings of the prior art, the application provides a recommendation method and a related device for investment products, so as to solve the problem that the existing mode cannot recommend the investment products which are most suitable for users to users.
In order to achieve the above object, the present application provides the following technical solutions:
the first aspect of the application provides a recommendation method for investment products, comprising the following steps:
acquiring user information of each user and product information of each investment product; wherein the user information includes personal basic information, financial information, and investment information;
constructing a knowledge graph between the user and the investment product based on the user information of each user and the product information of each investment product;
based on the preset personal basic information priority strategy, the financial information priority strategy and the investment information priority strategy, carrying out random walk on the knowledge graph to obtain a plurality of node sequences;
training each node sequence by applying a word vector model to obtain vectors of all nodes in each node sequence; wherein the nodes at least comprise user nodes and investment product nodes;
for each node sequence, calculating to obtain a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node by using a vector of each user node and a vector of each investment product node in the node sequence;
and generating a recommendation list of the investment products of each user according to the relevance score between each user and each investment product.
Optionally, in the above method for recommending an investment product, the performing random walk on the knowledge graph based on the preset personal basic information priority policy, the preset financial information priority policy and the preset investment information priority policy respectively to obtain a plurality of node sequences includes:
and carrying out random walk with a plurality of preset walk steps on the knowledge graph based on the preset personal basic information priority strategy, the financial information priority strategy and the investment information priority strategy respectively to obtain a plurality of first type node sequences biasing the personal basic information, a plurality of second type node sequences biasing the financial information and a plurality of third type node sequences biasing the investment information.
Optionally, in the above method for recommending an investment product, the calculating, for each node sequence, a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node by using a vector of each user node and a vector of each investment product node in the node sequence includes:
calculating the cosine similarity of the vector of each user node in the node sequence and the vector of each investment product node respectively aiming at each node sequence;
determining a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node according to the cosine similarity of the vector of each user node and the vector of each investment product node; the higher the cosine similarity is, the higher the determined relevance score is.
Optionally, in the above method for recommending investment products, the generating a recommendation list of investment products for each user according to a relevance score between each user and each investment product includes:
calculating average values of relevance aliquots between the user and each investment product respectively aiming at each user;
respectively sequencing the investment products according to the order of the average value from large to small for each user, and taking the investment products sequenced in the previous N times as a recommendation list of the investment products of the users; wherein N is a positive integer.
Optionally, in the above method for recommending investment products, after generating a recommendation list of investment products for each user according to the relevance score between each user and each investment product, the method further includes:
and monitoring state information of each investment product in the investment market in real time, and feeding back benefit expectations and risk reports of each investment product to each user at fixed time.
A second aspect of the present application provides an investment product recommending apparatus, comprising:
an acquisition unit configured to acquire user information of each user and product information of each investment product; wherein the user information includes personal basic information, financial information, and investment information;
a construction unit, configured to construct a knowledge graph between the user and the investment product based on user information of each user and product information of each investment product;
the wander unit is used for carrying out random wander on the knowledge graph based on the preset personal basic information priority strategy, the preset financial information priority strategy and the preset investment information priority strategy respectively to obtain a plurality of node sequences;
the training unit is used for training each node sequence by applying a word vector model to obtain the vector of each node in each node sequence; wherein the nodes at least comprise user nodes and investment product nodes;
the scoring unit is used for calculating and obtaining a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node by using the vector of each user node and the vector of each investment product node in the node sequence for each node sequence respectively;
and the generation unit is used for generating a recommendation list of the investment products of each user according to the relevance score between each user and each investment product.
Optionally, in the above investment product recommending apparatus, the walk-around unit includes:
and the wandering subunit is used for carrying out random wandering of a plurality of preset wandering steps on the knowledge graph based on the preset personal basic information priority strategy, the preset financial information priority strategy and the preset investment information priority strategy respectively to obtain a plurality of first type node sequences for biasing the personal basic information, a plurality of second type node sequences for biasing the financial information and a plurality of third type node sequences for biasing the investment information.
Optionally, in the above-mentioned investment product recommendation device, the scoring unit includes:
a first calculation unit configured to calculate, for each node sequence, a cosine similarity of a vector of each of the user nodes and a vector of each of the investment product nodes in the node sequence;
a determining unit, configured to determine a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node according to a cosine similarity of a vector of each user node and a vector of each investment product node; the higher the cosine similarity is, the higher the determined relevance score is.
Optionally, in the above investment product recommending apparatus, the generating unit includes:
a second calculation unit configured to calculate, for each of the users, an average value of respective correlations equally divided between the user and each of the investment products;
a ranking unit, configured to rank, for each of the users, the investment products in order from the average value to the small, and use the investment products ranked in the top N as a recommendation list of the investment products of the user; wherein N is a positive integer.
Optionally, in the above investment product recommending apparatus, the recommending apparatus further includes:
and the monitoring unit is used for monitoring the state information of each investment product in the investment market in real time and feeding back the benefit expectation and the risk report of each investment product to each user at regular time.
A third aspect of the present application provides an electronic device, comprising:
one or more processors;
a memory having one or more programs stored thereon;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of recommending investment products according to any of the preceding claims.
A fourth aspect of the present application provides a computer storage medium storing a program which, when executed, is adapted to carry out a method of recommending an investment product according to any one of the preceding claims.
The application provides a recommendation method of investment products, which comprises the steps of obtaining user information of each user and product information of each investment product; wherein the user information includes personal basic information, financial information, and investment information. And then, based on the user information of each user and the product information of each investment product, constructing a knowledge graph between the user and the investment product, and carrying out random walk on the knowledge graph based on a preset personal basic information priority strategy, a preset financial information priority strategy and a preset investment information priority strategy respectively to obtain a plurality of node sequences, thereby realizing the investment product meeting the user requirements in three aspects. Secondly, training each node sequence by using a word vector model to obtain vectors of all nodes in each node sequence, and calculating a relevance score between a user corresponding to each user node and an investment product corresponding to each investment product node by using the vectors of each user node and the vectors of each investment product node in each node sequence according to each node sequence, and finally generating a recommendation list of the investment products of each user according to the relevance score between each user and each investment product, thereby realizing a method capable of accurately recommending the investment product of the most suitable user for the user based on a knowledge graph.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a recommendation method for investment products according to an embodiment of the present application;
FIG. 2 is a flowchart of a method for calculating a relevance score according to another embodiment of the present disclosure;
FIG. 3 is a flow chart of a method of generating a recommendation list of investment products according to another embodiment of the present application;
FIG. 4 is a schematic structural view of a recommending apparatus for investment products according to another embodiment of the present application;
fig. 5 is a schematic structural diagram of a scoring unit according to another embodiment of the present disclosure;
fig. 6 is a schematic structural diagram of a generating unit according to another embodiment of the present application.
Detailed Description
The following description of the embodiments of the present invention 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 invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In this application, relational terms such as first and second, and the like are 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 terms "comprises," "comprising," or any other variation thereof, are 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. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The embodiment of the application provides a recommendation method of investment products, as shown in fig. 1, comprising the following steps:
s101, acquiring user information of each user and product information of each investment product, wherein the user information comprises personal basic information, financial information and investment information.
The personal basic information may specifically be basic information of a user such as name, age, job, international, marital, health status, and family condition. The financial information may include, in particular, financial information such as household income, personal assets, and liability status of the user. The investment information mainly comprises the risk preference, investment experience, historical investment condition of the user, namely the information of the investment products purchased by the user previously, and the like. The product information of the investment product mainly includes information about the investment product such as the type of the product, risk condition, threshold, expected income, etc.
S102, constructing a knowledge graph between the user and the investment product based on the user information of each user and the product information of each investment product.
Knowledge graph is essentially a knowledge base called semantic network, i.e. one with a directed graph structure. Wherein, the nodes of the knowledge graph represent entities or concepts, and the edges of the knowledge graph connecting the nodes represent various association relations between the entities or concepts, such as similarity relations between two entities, etc.
The knowledge graph to be constructed in the embodiment of the application is the knowledge graph between each user and the investment product, specifically, each user and the investment product are taken as the nodes in the knowledge graph, then the association relationship between the user and the investment product is determined according to the user information of the user and the product information of the investment product, and then the nodes of each user and the nodes of the investment product are connected according to the association relationship between the user and the investment product, so that the knowledge graph between the user and the investment product is formed. Namely, in the embodiment of the application, the constructed knowledge graph is as follows: g= (U, P, R), where U is an entity user set, P is an entity investment product set, and R represents a correlation attribute of a corresponding entity, that is, an association relationship between the two.
It should be noted that, the financial requirements of the user are closely related to the personal situation of the user, so the association relationship between the user and the investment product can be determined according to the user information of the user and the product information of the investment product. Specifically, the age and family status of the user are in different stages, the types of the biased investment products are different, for example, the user is in the family formation period, the income is continuously increased, and the expectation of the living standard is usually higher later, so that the demand of the user in the stage is biased to a mobile and high-benefit product with larger risk, and the association relationship between the user and the investment product conforming to the biased type of the user can be determined according to the personal basic information of the user and the investment information of the investment product. The economic condition of the user determines the acceptable investment threshold of the user, so that the user and the investment product meeting the investment threshold of the user can be determined according to the financial information of the user. The risk bearing capacity of the user, the personal preference is closely related to the investment preference of the user and the previous investment experience, so that the investment product which accords with the risk bearing capacity of the user can be determined according to the investment information of each user. Therefore, a knowledge graph between the user and the investment product can be constructed according to the user information of the user and the product information of the investment product.
S103, performing random walk on the knowledge graph based on a preset personal basic information priority strategy, a preset financial information priority strategy and a preset investment information priority strategy respectively to obtain a plurality of node sequences.
It should be noted that, the step of performing the walk on the knowledge graph may be colloquially understood as selecting an initial node from all nodes of the knowledge graph, then selecting a node from all nodes connected with the initial node, then selecting a node from all nodes connected with the node selected in the previous step, and sequentially selecting the nodes downwards until the set end condition is met, so that the selected nodes form an ordered node sequence. Specifically, in the embodiment of the present application, the corresponding node of each user is selected as the starting node during the random walk to perform the random walk, so that the recommendation list of each user can be obtained finally.
The personal basic information priority strategy refers to a random walk strategy for biasing personal basic information, and is generally understood as that when one path is selected from a plurality of paths connected with one target node so as to determine the next node, if one node is connected with the target node based on the personal basic information, the probability of selecting the node is larger than the probability of selecting other nodes, or the node is directly selected as the next node. Similarly, the financial information priority policy is a random walk policy that biases the financial information. The investment information priority policy is a random walk policy that biases the investment information.
Because the type of the investment product suitable for the user can be analyzed according to the personal basic information of the user, the financial information of the user feels the investment threshold of the investment product acceptable by the user, and the investment information of the user is closely related to the risk bearing capacity of the user, the personal basic information priority strategy, the financial information priority strategy and the investment information priority strategy are adopted in the embodiment of the application, and random walk is carried out on the knowledge graph to obtain a plurality of node sequences for biasing the personal basic information, the financial information and the investment information, so that the investment product in the finally obtained recommendation list is obtained based on the personal basic information, the financial information and the investment information.
Optionally, in another embodiment of the present application, a specific implementation manner of step S103 specifically includes: and carrying out random walk with a plurality of preset walk steps on the knowledge graph based on a preset personal basic information priority strategy, a financial information priority strategy and an investment information priority strategy respectively to obtain a plurality of first type node sequences for biasing the personal basic information, a plurality of second type node sequences for biasing the financial information and a plurality of third type node sequences for biasing the investment information.
Specifically, in the embodiment of the application, when the personal basic information priority strategy, the financial information priority strategy and the investment information priority strategy are respectively adopted for random walk, the walk is carried out for a plurality of times, the number of steps of each walk is preset walk deployment, so that a plurality of node sequences with the same length are obtained based on each walk strategy, the node sequences are processed subsequently to obtain a recommendation list of users, the occurrence of contingency can be greatly avoided, and the recommended investment products more meet the requirements of the users.
S104, training each node sequence by applying a word vector model to obtain the vector of each node in each node sequence.
Wherein the nodes include at least a user node and an investment product node.
Alternatively, in the application embodiment, the Word vector model Word2vec may be used to train each node sequence accordingly, so as to obtain the vector of each node in each node sequence. Since the Word vector model Word2vec can be used to map each Word to a vector and to represent Word-to-Word relationships. Therefore, the same node in different node sequences, the obtained node vectors are different, namely, the obtained node vectors are influenced by the relationship with other nodes in the node sequences, so that the relationship between the user and the investment product can be reflected according to the node vectors.
S105, calculating a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node by using the vector of each user node and the vector of each investment product node in the node sequence for each node sequence.
Because the higher the similarity between the vector of the user node and the vector of the investment product node, the more the investment product corresponding to the investment product node meets the requirement of the user corresponding to the user node, the similarity between the vector of each user node and the vector of each investment product node in the node sequence can be determined specifically for each node sequence, and then the relevance score between the user corresponding to each user node in the node sequence and the investment product corresponding to each investment product node in the node sequence is determined according to the similarity between the node vectors.
Optionally, in another embodiment of the present application, a specific implementation manner of step S105, as shown in fig. 2, includes:
s201, respectively calculating cosine similarity of vectors of each user node and vectors of each investment product node in the node sequence aiming at each node sequence.
It should be noted that, the vector obtained by training the Word through the Word vector model Word2vec is a distributed Word vector, and the included angle between the distributed Word vectors can represent the correlation between the words, and because the included angle between the vectors can be obtained by cosine theorem calculation, in the embodiment of the present application, the cosine value of the vector of each user node and the cosine value of the vector of each investment product node in the node sequence are calculated for each node sequence respectively, and then the cosine similarity of the vector of each user node and the cosine of the vector of each investment product node is determined according to the cosine value.
S202, determining a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node according to the cosine similarity of the vector of each user node and the vector of each investment product node.
The higher the cosine similarity, the higher the determined relevance score.
Alternatively, a corresponding relationship of cosine similarity and relevance equal division may be pre-established, and then, a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node may be determined subsequently according to the corresponding relationship of cosine similarity and relevance equal division. Alternatively, the cosine similarity between the vector of the user node and the vector of the investment product node is substituted into the mapping formula through a preset mapping formula, so as to obtain the relevance equal score of the investment product corresponding to the vector of the user node and the vector of the investment product node. Of course, other manners may be adopted, for example, the cosine similarity between the user corresponding to the vector of the user node and the vector of the investment product node is directly determined as the relevance score between the user corresponding to the vector of the user node and the vector of the investment product corresponding to the vector of the investment product node, which shall fall into the protection scope of the present application.
In the prior art, the score between the user and the investment product is calculated, that is, several features, such as score, sales, etc., are usually found, then a simple linear function is found, the weight value corresponding to each feature in the linear function is manually adjusted, and then the score between the user and each investment product is calculated through the linear function. This method is very simple but suffers from a number of problems: the weights of all features are adjusted every time one feature is added; when the characteristics are more, the manual weight adjustment mode cannot be completed, and a better solution is difficult to find; along with the continuous change of data, the weight needs to be manually readjusted at intervals, and the overall manual adjustment is labor-consuming and has poor effect.
S106, generating a recommendation list of the investment products of each user according to the relevance scores between each user and each investment product.
After obtaining the relevance score between each user and each investment product, a recommendation list of the investment products of each user can be generated by adopting different recommendation strategies based on the relevance score between each user and each investment product, and the recommendation list is recommended to the user. Specifically, from user nodes and investment product nodes in a sequence obtained by wandering off personal basic information, financial information and investment information, respectively, a plurality of investment products with highest relevance scores to target users in the node sequence are selected, and then the selected three investment products are combined to obtain a recommendation list containing investment product combinations. Of course, the relevance scores of the users and the investment products obtained by the previous calculation can be ranked in order from the big to the small for each user, so as to obtain a recommendation list of the investment products of the users. Or calculating the rank of the investment product through a ranking algorithm, and then generating a recommendation list and the like according to the rank of the investment product.
Optionally, in another embodiment of the present application, a specific implementation manner of step S106, as shown in fig. 3, specifically includes:
s301, calculating average values of relevance aliquots between the users and each investment product for each user.
Because the knowledge graph is traversed for multiple times, the same user node and the same investment product may appear on different node sequences at the same time, and thus multiple relevance scores exist for the same user and the same investment product. Thus, in the present embodiment, for each user, an average of all the correlations between that user and each investment product is calculated. If there is only one relevance score between the user and a certain investment product, the relevance score between the user and the investment product is the average value of all relevance scores between the user and the investment product, so that calculation of the average value may not be continued.
S302, respectively ordering the investment products according to the order from the average value to the small value for each user, taking the investment products ordered in the first N as a recommendation list of the investment products of the users, wherein N is a positive integer.
Optionally, in another embodiment of the present application, after performing step S106, the method may further include: the status information of each investment product in the investment market is monitored in real time, and the benefit expectations and the risk reports of each investment product are fed back to each user at regular time.
Specifically, an investment market monitor of the investment product is established to detect the state of the investment product, and a benefit expectation and a risk report are generated according to the actual situation of the investment market. The analysis report may include, among other things, risk monitoring information, purchase advice, and the like. And the risk report may be fed back to the user in the form of daily, weekly, and monthly reports.
According to the recommendation method of the investment products, user information of each user and product information of each investment product are obtained; wherein the user information includes personal basic information, financial information, and investment information. And then, based on the user information of each user and the product information of each investment product, constructing a knowledge graph between the user and the investment product, and carrying out random walk on the knowledge graph based on a preset personal basic information priority strategy, a preset financial information priority strategy and a preset investment information priority strategy respectively to obtain a plurality of node sequences, thereby realizing the investment product meeting the user requirements in three aspects. Secondly, training each node sequence by using a word vector model to obtain vectors of all nodes in each node sequence, and calculating a relevance score between a user corresponding to each user node and an investment product corresponding to each investment product node by using the vectors of each user node and the vectors of each investment product node in each node sequence according to each node sequence, so that a recommendation list of the investment product of each user is finally generated according to the relevance score between each user and each investment product, and a method capable of accurately recommending proper investment products for users is realized based on a knowledge graph.
Another embodiment of the present application provides a recommendation device for investment products, as shown in fig. 4, including:
an acquisition unit 401 for acquiring user information of each user and product information of each investment product.
Wherein the user information includes personal basic information, financial information, and investment information.
A construction unit 402, configured to construct a knowledge graph between the user and the investment product based on the user information of each user and the product information of each investment product.
The walk unit 403 is configured to perform random walk on the knowledge graph based on a preset personal basic information priority policy, a preset financial information priority policy, and a preset investment information priority policy, respectively, to obtain a plurality of node sequences.
The training unit 404 is configured to train each node sequence by applying a word vector model to obtain a vector of each node in each node sequence; wherein the nodes include at least user nodes and investment product nodes.
And the scoring unit 405 is configured to calculate, for each node sequence, a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node by using the vector of each user node and the vector of each investment product node in the node sequence.
A generating unit 406, configured to generate a recommendation list of investment products for each user according to the relevance score between each user and each investment product.
It should be noted that, the specific working process of the above units in the embodiments of the present application may refer to the steps S101 to S106 in the embodiments of the above method, which is not described herein again.
Optionally, in the investment product recommending apparatus provided in another embodiment of the present application, the wandering unit 403 includes:
the wandering subunit is used for performing random wandering of a plurality of preset wandering steps on the knowledge graph based on a preset personal basic information priority strategy, a preset financial information priority strategy and a preset investment information priority strategy respectively to obtain a plurality of first type node sequences for biasing the personal basic information, a plurality of second type node sequences for biasing the financial information and a plurality of third type node sequences for biasing the investment information.
Optionally, in the recommending apparatus for investment products according to another embodiment of the present application, as shown in fig. 5, the scoring unit 405 specifically includes:
a first calculation unit 501 is configured to calculate, for each node sequence, a cosine similarity of a vector of each user node in the node sequence and a vector of each investment product node, respectively.
A determining unit 502, configured to determine a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node according to the cosine similarity of the vector of each user node and the vector of each investment product node.
The higher the cosine similarity, the higher the determined relevance score.
It should be noted that, the specific working processes of the first calculating unit 501 and the determining unit 502 may refer to the step S201 and the step S202 in the above method embodiment, which are not described herein.
Optionally, in the recommending apparatus for investment products according to another embodiment of the present application, the generating unit 406, as shown in fig. 6, specifically includes:
a second calculation unit 601 is configured to calculate, for each user, an average value of respective relevance aliquots between the user and each investment product.
A ranking unit 602, configured to rank the investment products according to the order from the average value to the small for each user, and take the investment products ranked in the top N number as a recommendation list of the investment products of the users, where N is a positive integer.
It should be noted that, the specific working processes of the second calculating unit 601 and the sorting unit 602 may refer to the step S301 and the step S302 in the above method embodiments, which are not described herein.
Optionally, in the above investment product recommending apparatus, the recommending apparatus further includes:
and the monitoring unit is used for monitoring the state information of each investment product in the investment market in real time and feeding back the benefit expectation and the risk report of each investment product to each user at regular time.
According to the recommendation device for the investment products, the user information of each user and the product information of each investment product are obtained through the obtaining unit; wherein the user information includes personal basic information, financial information, and investment information. Then the construction unit constructs a knowledge graph between the user and the investment product based on the user information of each user and the product information of each investment product, and the wandering unit performs random wander on the knowledge graph based on a preset personal basic information priority strategy, a preset financial information priority strategy and a preset investment information priority strategy respectively to obtain a plurality of node sequences, so that the investment product meeting the user requirements is realized in three aspects. Secondly, the training unit trains each node sequence by applying a word vector model to obtain vectors of all nodes in each node sequence, calculates and obtains a relevance score between a user corresponding to each user node and an investment product corresponding to each investment product node by using the vectors of each user node and the vectors of each investment product node in the node sequence for each node sequence, and finally the generating unit generates a recommendation list of the investment products of each user according to the relevance score between each user and each investment product, so that a recommendation device of the investment products based on a knowledge graph is realized, and suitable recommendation can be accurately realized for the user.
Another embodiment of the present application provides an electronic device, including:
one or more processors, memory.
Wherein the memory has one or more programs stored thereon, which when executed by the one or more processors cause the one or more processors to implement the method of recommending investment products as provided by any of the method embodiments described above.
Another embodiment of the present application further provides a computer storage medium for storing a program that, when executed, is configured to implement the investment product recommendation method provided in any one of the method embodiments described above.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of functionality in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. 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 invention. Thus, the present invention 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 (8)
1. A method of recommending investment products, comprising:
acquiring user information of each user and product information of each investment product; wherein the user information includes personal basic information, financial information, and investment information; the investment information includes information of investment products previously purchased by the user;
based on the user information of each user and the product information of each investment product, each user and each investment product are used as nodes in a knowledge graph, the association relation between the user and the investment product is determined according to the user information of the user and the product information of the investment product, and then the nodes of each user and the nodes of the investment product are connected according to the association relation between the user and the investment product, so that the knowledge graph between the user and the investment product is constructed;
based on the preset personal basic information priority strategy, the financial information priority strategy and the investment information priority strategy, carrying out random walk with a plurality of preset walk steps on the knowledge graph to obtain a plurality of first type node sequences biasing the personal basic information, a plurality of second type node sequences biasing the financial information and a plurality of third type node sequences biasing the investment information;
training each node sequence by applying a word vector model to obtain vectors of all nodes in each node sequence; wherein the nodes at least comprise user nodes and investment product nodes;
for each node sequence, calculating to obtain a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node by using a vector of each user node and a vector of each investment product node in the node sequence;
and generating a recommendation list of the investment products of each user according to the relevance score between each user and each investment product.
2. The method according to claim 1, wherein for each node sequence, calculating, using the vector of each user node and the vector of each investment product node in the node sequence, a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node, includes:
calculating the cosine similarity of the vector of each user node in the node sequence and the vector of each investment product node respectively aiming at each node sequence;
determining a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node according to the cosine similarity of the vector of each user node and the vector of each investment product node; the higher the cosine similarity is, the higher the determined relevance score is.
3. The method of claim 1, wherein generating a recommended list of investment products for each of the users based on the relevance scores between each of the users and each of the investment products comprises:
calculating average values of relevance aliquots between the user and each investment product respectively aiming at each user;
respectively sequencing the investment products according to the order of the average value from large to small for each user, and taking the investment products sequenced in the previous N times as a recommendation list of the investment products of the users; wherein N is a positive integer.
4. The method of claim 1, wherein generating a recommended list of investment products for each of the users based on the relevance scores between each of the users and each of the investment products further comprises:
and monitoring state information of each investment product in the investment market in real time, and feeding back benefit expectations and risk reports of each investment product to each user at fixed time.
5. A recommendation device for investment products, comprising:
an acquisition unit configured to acquire user information of each user and product information of each investment product; wherein the user information includes personal basic information, financial information, and investment information; the investment information includes information of investment products previously purchased by the user;
the construction unit is used for taking each user and each investment product as a node in the knowledge graph based on the user information of each user and the product information of each investment product, determining the association relationship between the user and the investment product according to the user information of the user and the product information of the investment product, and further connecting the nodes of each user and the nodes of the investment product according to the association relationship between the user and the investment product to construct the knowledge graph between the user and the investment product;
the wander unit is used for carrying out random wander on the knowledge graph based on the preset personal basic information priority strategy, the preset financial information priority strategy and the preset investment information priority strategy respectively to obtain a plurality of node sequences;
the training unit is used for training each node sequence by applying a word vector model to obtain the vector of each node in each node sequence; wherein the nodes at least comprise user nodes and investment product nodes;
the scoring unit is used for calculating and obtaining a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node by using the vector of each user node and the vector of each investment product node in the node sequence for each node sequence respectively;
a generating unit, configured to generate a recommendation list of investment products of each user according to a relevance score between each user and each investment product;
the walk unit includes:
and the wandering subunit is used for carrying out random wandering of a plurality of preset wandering steps on the knowledge graph based on the preset personal basic information priority strategy, the preset financial information priority strategy and the preset investment information priority strategy respectively to obtain a plurality of first type node sequences for biasing the personal basic information, a plurality of second type node sequences for biasing the financial information and a plurality of third type node sequences for biasing the investment information.
6. The apparatus of claim 5, wherein the scoring unit comprises:
a first calculation unit configured to calculate, for each node sequence, a cosine similarity of a vector of each of the user nodes and a vector of each of the investment product nodes in the node sequence;
a determining unit, configured to determine a relevance score between the user corresponding to each user node and the investment product corresponding to each investment product node according to a cosine similarity of a vector of each user node and a vector of each investment product node; the higher the cosine similarity is, the higher the determined relevance score is.
7. An electronic device, comprising:
one or more processors;
a memory having one or more programs stored thereon;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the investment product recommendation method of any one of claims 1 to 4.
8. A computer storage medium storing a program which, when executed, is adapted to carry out the investment product recommendation method according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010392822.4A CN111581516B (en) | 2020-05-11 | 2020-05-11 | Investment product recommending method and related device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010392822.4A CN111581516B (en) | 2020-05-11 | 2020-05-11 | Investment product recommending method and related device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111581516A CN111581516A (en) | 2020-08-25 |
| CN111581516B true CN111581516B (en) | 2024-04-12 |
Family
ID=72124864
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010392822.4A Active CN111581516B (en) | 2020-05-11 | 2020-05-11 | Investment product recommending method and related device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111581516B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112035681B (en) * | 2020-09-17 | 2023-10-24 | 中国银行股份有限公司 | Method and device for determining credit card rate information based on knowledge graph |
| CN112102097A (en) * | 2020-09-28 | 2020-12-18 | 中国银行股份有限公司 | Investment capacity evaluation method and device, electronic equipment and computer storage medium |
| CN112417167A (en) * | 2020-11-20 | 2021-02-26 | 中国平安人寿保险股份有限公司 | Construction method and device of insurance knowledge graph, computer equipment and storage medium |
| CN112925917A (en) * | 2021-02-25 | 2021-06-08 | 中国平安人寿保险股份有限公司 | Method and device for constructing product knowledge graph, terminal and storage medium |
| CN113159972B (en) * | 2021-05-20 | 2024-05-24 | 深圳前海微众银行股份有限公司 | Combination determination method, device, electronic equipment and computer readable storage medium |
| CN114022270B (en) * | 2021-10-29 | 2024-05-24 | 平安银行股份有限公司 | Asset data processing method, related equipment and medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107798608A (en) * | 2017-10-19 | 2018-03-13 | 深圳市耐飞科技有限公司 | A kind of investment product combined recommendation method and system |
| CN108920527A (en) * | 2018-06-07 | 2018-11-30 | 桂林电子科技大学 | A kind of personalized recommendation method of knowledge based map |
| CN110223155A (en) * | 2019-04-25 | 2019-09-10 | 深圳壹账通智能科技有限公司 | Method for pushing, device and the computer equipment of investment recommendation information |
| WO2020048051A1 (en) * | 2018-09-04 | 2020-03-12 | 深圳壹账通智能科技有限公司 | Financial product recommendation method, server and computer readable storage medium |
| WO2020070463A1 (en) * | 2018-10-03 | 2020-04-09 | Standard Life Employee Services Limited | System for modelling financial information |
-
2020
- 2020-05-11 CN CN202010392822.4A patent/CN111581516B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107798608A (en) * | 2017-10-19 | 2018-03-13 | 深圳市耐飞科技有限公司 | A kind of investment product combined recommendation method and system |
| CN108920527A (en) * | 2018-06-07 | 2018-11-30 | 桂林电子科技大学 | A kind of personalized recommendation method of knowledge based map |
| WO2020048051A1 (en) * | 2018-09-04 | 2020-03-12 | 深圳壹账通智能科技有限公司 | Financial product recommendation method, server and computer readable storage medium |
| WO2020070463A1 (en) * | 2018-10-03 | 2020-04-09 | Standard Life Employee Services Limited | System for modelling financial information |
| CN110223155A (en) * | 2019-04-25 | 2019-09-10 | 深圳壹账通智能科技有限公司 | Method for pushing, device and the computer equipment of investment recommendation information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111581516A (en) | 2020-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111581516B (en) | Investment product recommending method and related device | |
| US9223900B2 (en) | Machine optimization devices, methods, and systems | |
| Capuano et al. | Fuzzy group decision making for influence-aware recommendations | |
| WO2018126953A1 (en) | Seed population expanding method, device, information releasing system and storing medium | |
| CN110796477A (en) | Advertisement display method and device, electronic equipment and readable storage medium | |
| Jadhav et al. | Analytic hierarchy process (AHP), weighted scoring method (WSM), and hybrid knowledge based system (HKBS) for software selection: a comparative study | |
| US20110196716A1 (en) | Lead qualification based on contact relationships and customer experience | |
| US20120143718A1 (en) | Optimization of a web-based recommendation system | |
| Hashemian et al. | A fuzzy hybrid group decision support system approach for the supplier evaluation process | |
| Rodríguez et al. | Using linguistic incomplete preference relations to cold start recommendations | |
| EP2960849A1 (en) | Method and system for recommending an item to a user | |
| WO2009094672A2 (en) | Belief propagation for generalized matching | |
| US9489638B2 (en) | Method and apparatus for propagating user preference information in a communications network | |
| US20110093457A1 (en) | Method for calculating resource points of resource information and distributing points | |
| CN108460082A (en) | A kind of recommendation method and device, electronic equipment | |
| JP6819355B2 (en) | Recommendation generation | |
| CN109242654A (en) | A kind of item recommendation method and system | |
| CN109711871A (en) | A kind of potential customers determine method, apparatus, server and readable storage medium storing program for executing | |
| Naamani Dery et al. | Iterative voting under uncertainty for group recommender systems | |
| Mikeli et al. | A multi-criteria recommendation method for interval scaled ratings | |
| US20140257980A1 (en) | Displaying promotion information | |
| US8843428B2 (en) | Survey prioritization engine | |
| CN112163886A (en) | Real-time bidding advertisement resource allocation method based on reinforcement learning | |
| WO2002079901A2 (en) | Customer preference system | |
| Gerogiannis et al. | Personalised fuzzy recommendation for high involvement products |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |