Personalized credit product recommendation system and method based on user portrait

Abstract

The invention relates to the technical field of credit product recommendation, in particular to a personalized credit product recommendation system and method based on a user portrait. The method comprises the following steps: obtaining user basic information data and user behavior data, generating a user credit qualification value based on the user basic information data, and examining the credit qualification of a user based on the user credit qualification value; obtaining user basic information data and user behavior data with credit qualification, and performing user portrait construction based on the user basic information data and the user behavior data to obtain user credit intention features; and matching credit products based on the credit intention features of the user, and recommending the credit products to the user. The user experience can be improved, the credit service efficiency is improved, and the credit risk is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of credit product recommendation, and in particular to a personalized credit product recommendation system and method based on user portraits. Background Art

[0002] With the rapid development of Internet finance, the credit market has shown an unprecedented prosperity. However, faced with a large number of diverse credit products, how to accurately identify users' credit needs and recommend suitable products to them has become an important issue that financial institutions need to solve urgently.

[0003] User portrait is a comprehensive description of user attributes, behaviors, preferences and other information, and is the basis of personalized recommendation system. In the field of credit, the construction of user portrait needs to consider multiple dimensions of user information, including but not limited to: age, gender, occupation, income, etc. The user's past credit application record, credit limit usage, repayment record, etc. are important bases for evaluating the user's credit status. Browsing and search behavior: The user's browsing and searching behavior of credit products on the Internet platform can directly reflect their current credit needs and preferences.

[0004] Traditional credit product recommendation methods often rely on manual experience and simple rules, which is not only inefficient but also difficult to meet the personalized and diversified needs of users.

[0005] Therefore, a personalized credit product recommendation system based on user portraits came into being and became an effective means to solve this problem. Summary of the invention

[0006] The purpose of the present invention is to provide a personalized credit product recommendation system and method based on user portraits: to solve the technical problem that the existing solutions often rely on manual experience and simple rules in the process of recommending user credit products, which is not only inefficient but also difficult to meet the personalized and diversified needs of users.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] On the one hand, the personalized credit product recommendation system based on user portraits includes a user data acquisition module, a user credit qualification review module, a user portrait construction module and a credit product matching module;

[0009] The user data acquisition module is used to acquire user basic information data and user behavior data, wherein the user basic information data includes age, education, social security record information and information on applied credit lines, and the user behavior data includes the number of times the user clicks on credit products, the number of times the webpage where the credit products are located is browsed, and the keywords that the user searches for credit products;

[0010] The user credit qualification review module is used to generate a user credit qualification value based on the user's basic information data, and review the user's credit qualification based on the user's credit qualification value;

[0011] The user profile building module is used to obtain basic information data and user behavior data of users with credit qualifications, and to build user profiles based on the basic information data and user behavior data to obtain user credit intention characteristics;

[0012] The credit product matching module is used to match credit products based on user credit intention characteristics and recommend credit products to users.

[0013] Furthermore, generating a user credit qualification value based on the user basic information data includes the following process:

[0014] Get the user's age A based on the user's basic information data;

[0015] Based on the user's basic information data, the user's education level is obtained, and the user's corresponding credit qualification basic value B is queried based on the user's education level. Among them, the credit qualification basic value corresponding to a high school education level is 10, the credit qualification basic value corresponding to a bachelor's degree is 20, and the credit qualification basic value corresponding to a postgraduate degree is 30;

[0016] Based on the user's basic information data, obtain the user's social security record information, and based on the social security record information, obtain the number of consecutive social security payment months C of the user;

[0017] Based on the basic information data of the user, the information on the credit line applied for is obtained, and based on the information on the credit line applied for, the total amount of credit applied for by the user D is obtained;

[0018] Substitute age A, credit qualification base value B, consecutive social security payment months C, and user application credit amount D into the user credit qualification value calculation formula to obtain the user credit qualification value, which is as follows:

[0019]

[0020] Among them, LMS is the user's credit qualification value.

[0021] Furthermore, reviewing the credit qualification of the user based on the user's credit qualification value includes the following process:

[0022] A credit qualification value threshold is obtained, wherein the credit qualification value threshold is stored in the system and is set by the system to determine whether the user's credit qualification value exceeds the credit qualification value threshold. If so, the user is credit qualified, and if not, the user is not credit qualified.

[0023] Furthermore, constructing a user profile based on the user's basic information data and user behavior data to obtain the user's credit intention characteristics includes the following process:

[0024] The user's basic information data and user behavior data are recorded as text, and useless characters, segmentation words and stop words are removed from the text. The text is segmented using the Jieba Chinese segmentation component to obtain a word set of the text, where the words include age, education, social security, personal tax payment, click, browse and search;

[0025] Train the word set of the text based on the pre-trained word2vec model, and then use the trained model to generate word vectors;

[0026] The word vectors of the text are weighted averaged by the TF-IDF weighted statistical method to generate an average word vector to represent the text features. The final text feature vector is TF = {t 1 , t 2 ...t k}, TF={t 1 , t 2 ...t k} is recorded as the user's credit intention feature.

[0027] Furthermore, the word vectors of the text are weighted averaged by the TF-IDF weighted statistical method to generate an average word vector to represent the text features, which includes the following process:

[0028]

[0029] TF-IDF=TF w ×IDF;

[0030] The weights are obtained based on the TF-IDF value. Common words are given smaller weights, while less common words are given larger weights. That is, key feature words are extracted through weighted statistics.

[0031] Among them, f(w) represents the number of times word w appears in the text, TF w is the word frequency, It indicates the number of times other words appear in the text, n is the number of types of other words, and the value of i is 1, 2...n; dfw represents the number of texts in which word w appears.

[0032] Furthermore, matching credit products based on user credit intention characteristics and recommending credit products to users includes the following processes:

[0033] Use a neural network model to divide the dataset into training and testing sets;

[0034] Use the training set data to train the model and adjust the model parameters through cross-validation and other methods;

[0035] Monitor the loss function during training and the performance metrics on the validation set to avoid overfitting;

[0036] Use the test set to evaluate the performance of the model. Common evaluation indicators include accuracy, recall, F1 score, and AUC-ROC.

[0037] Tune the model based on the evaluation results, including adjusting model parameters and trying different model structures or algorithms;

[0038] Predict users’ credit intentions and preferences based on the trained model;

[0039] Based on the prediction results, the credit products that best meet the needs of users are recommended to them. The personalized recommendation algorithm includes a collaborative filtering algorithm, which combines the user's historical behavior and the behavior of similar users to recommend products.

[0040] On the other hand, a personalized credit product recommendation method based on user portraits includes:

[0041] Obtaining basic user information and behavior data, where basic user information includes age, education, social security record information, and individual tax payment information; and user behavior data includes the number of times a user clicks on a credit product, the number of times a user browses the webpage where the credit product is located, and the keywords that the user searches for the credit product;

[0042] Generate a user credit qualification value based on the user's basic information data, and review the user's credit qualification based on the user's credit qualification value;

[0043] Obtain basic information data and user behavior data of users with credit qualifications, and construct user portraits based on the basic information data and user behavior data to obtain user credit intention characteristics;

[0044] Match credit products based on user credit intention characteristics and recommend credit products to users.

[0045] Compared with the existing solutions, the present invention achieves the following beneficial effects:

[0046] Improve user experience: By accurately analyzing users' credit needs and preferences, the system can recommend credit products that best meet their personalized needs. This tailored recommendation method not only saves users time in screening and comparing products, but also improves user satisfaction and loyalty.

[0047] Improve credit service efficiency: Traditional credit product recommendation methods often rely on manual review and recommendation, which is inefficient and prone to errors. The recommendation system based on user portraits can automatically match users and products, greatly improving the efficiency and accuracy of credit services. At the same time, the system can also monitor user behavior and market changes in real time, adjust recommendation strategies in a timely manner, and ensure the timeliness and effectiveness of recommendations.

[0048] Reduce credit risk: Through in-depth analysis of users' credit status, repayment ability, etc., the system can more accurately assess users' credit risk. This helps financial institutions to be more cautious and prudent when recommending products, reduce non-performing loan rates, and improve overall risk control levels. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0050] Figure 1 It is a system block diagram of a personalized credit product recommendation system based on user portraits according to an embodiment of the present invention;

[0051] Figure 2 This is a workflow diagram of a personalized credit product recommendation system based on user portraits according to an embodiment of the present invention;

[0052] Figure 3 It is a workflow diagram of a personalized credit product recommendation method based on user portrait according to an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0055] This embodiment provides a personalized credit product recommendation system based on user portraits. Figure 1 is a system block diagram of a personalized credit product recommendation system based on user portraits according to an embodiment of the present invention. Figure 1 As shown, the system includes a user data acquisition module, a user credit qualification review module, a user portrait construction module and a credit product matching module;

[0056] The user data acquisition module is used to acquire user basic information data and user behavior data, wherein the user basic information data includes age, education, social security record information and information on applied credit lines, and the user behavior data includes the number of times the user clicks on credit products, the number of times the webpage where the credit products are located is browsed, and the keywords that the user searches for credit products;

[0057] The user credit qualification review module is used to generate a user credit qualification value based on the user's basic information data, and review the user's credit qualification based on the user's credit qualification value;

[0058] The user profile building module is used to obtain basic information data and user behavior data of users with credit qualifications, and to build user profiles based on the basic information data and user behavior data to obtain user credit intention characteristics;

[0059] The credit product matching module is used to match credit products based on user credit intention characteristics and recommend credit products to users.

[0060] In summary, the present invention obtains user basic information data and user behavior data, generates a user credit qualification value based on the user basic information data, and examines the user's credit qualification based on the user credit qualification value; obtains the basic information data and user behavior data of users with credit qualifications, and constructs a user portrait based on the basic information data and user behavior data to obtain the user's credit intention characteristics; matches credit products based on the user's credit intention characteristics, and recommends credit products to users. The present invention can improve user experience, improve credit service efficiency, and reduce credit risks.

[0061] In some embodiments, Figure 2is a workflow diagram of a personalized credit product recommendation system based on user portraits according to an embodiment of the present invention. Figure 2 As shown, generating a user credit qualification value based on user basic information data includes the following steps:

[0062] Step S201: Acquire the user's age A based on the user's basic information data;

[0063] Step S202: Based on the user's basic information data, the user's education level is obtained, and the credit qualification basic value B corresponding to the user is queried based on the user's education level, wherein the credit qualification basic value corresponding to a high school education level is 10, the credit qualification basic value corresponding to a bachelor's degree level is 20, and the credit qualification basic value corresponding to a postgraduate degree level is 30;

[0064] Step S203: obtaining the user's social security record information based on the user's basic information data, and obtaining the number of consecutive social security payment months C of the user based on the social security record information;

[0065] Step S204: obtaining information on the credit line applied for based on the basic information data of the user, and obtaining a total amount of credit applied for by the user D based on the information on the credit line applied for;

[0066] Step S205: Substitute age A, credit qualification base value B, consecutive social security payment months C and user application credit amount D into the user credit qualification value calculation formula to obtain the user credit qualification value, which is as follows:

[0067]

[0068] Among them, LMS is the user's credit qualification value.

[0069] In some embodiments, reviewing the credit qualification of a user based on the user's credit qualification value includes the following process:

[0070] A credit qualification value threshold is obtained, wherein the credit qualification value threshold is stored in the system and is set by the system to determine whether the user's credit qualification value exceeds the credit qualification value threshold. If so, the user is credit qualified, and if not, the user is not credit qualified.

[0071] In some embodiments, constructing a user profile based on user basic information data and user behavior data to obtain user credit intention characteristics includes the following process:

[0072] The user's basic information data and user behavior data are recorded as text, and useless characters, segmentation words and stop words are removed from the text. The text is segmented using the Jieba Chinese segmentation component to obtain a word set of the text, where the words include age, education, social security, personal tax payment, click, browse and search;

[0073] Train the word set of the text based on the pre-trained word2vec model, and then use the trained model to generate word vectors;

[0074] The word vectors of the text are weighted averaged by the TF-IDF weighted statistical method to generate an average word vector to represent the text features. The final text feature vector is TF = {t 1 , t 2 ...t k}, TF={t 1 , t 2 ...t k} is recorded as the user's credit intention feature.

[0075] Furthermore, the word vectors of the text are weighted averaged by the TF-IDF weighted statistical method to generate an average word vector to represent the text features, which includes the following process:

[0076]

[0077] TF-IDF=TF w ×IDF;

[0078] The weights are obtained based on the TF-IDF value. Common words are given smaller weights, while less common words are given larger weights. That is, key feature words are extracted through weighted statistics.

[0079] Among them, f(w) represents the number of times word w appears in the text, TF w is the word frequency, It indicates the number of times other words appear in the text, n is the number of types of other words, and the value of i is 1, 2...n; dfw represents the number of texts in which word w appears.

[0080] Furthermore, the dataset is divided into a training set and a test set using a neural network model;

[0081] Use the training set data to train the model and adjust the model parameters through cross-validation and other methods;

[0082] Monitor the loss function during training and the performance metrics on the validation set to avoid overfitting;

[0083] Use the test set to evaluate the performance of the model. Common evaluation indicators include accuracy, recall, F1 score, and AUC-ROC.

[0084] Tune the model based on the evaluation results, including adjusting model parameters and trying different model structures or algorithms;

[0085] Predict users’ credit intentions and preferences based on the trained model;

[0086] Based on the prediction results, the credit products that best meet the needs of users are recommended to them. The personalized recommendation algorithm includes a collaborative filtering algorithm, which combines the user's historical behavior and the behavior of similar users to recommend products.

[0087] It is worth mentioning that common recommendation algorithms include collaborative filtering, content-based recommendation, hybrid recommendation, etc.

[0088] Collaborative filtering: recommends products by analyzing the similarities between users and products. In the credit field, you can use information such as a user's credit history and consumer behavior to find other users who are similar to them, and then recommend products based on their credit product usage.

[0089] Content-based recommendation: Based on the characteristics and needs of the user portrait, directly filter out qualified products from the product library for recommendation. This method is more intuitive, but it requires an accurate description of product characteristics and user needs.

[0090] Hybrid recommendation: Combining the advantages of collaborative filtering and content-based recommendation, comprehensively considering the similarities between users and products, as well as the matching degree between users and products, to improve the accuracy and diversity of recommendations.

[0091] In practical applications, the recommendation algorithm also needs to be optimized, such as through machine learning or deep learning algorithms to improve the accuracy and efficiency of recommendations. For example, algorithms such as BP neural network and K-means clustering algorithm can be used to further analyze and process user portraits and product features to improve the recommendation effect.

[0092] In some embodiments, the present invention further provides a personalized credit product recommendation method based on user portraits. Figure 3 is a workflow diagram of a personalized credit product recommendation method based on user portraits according to an embodiment of the present invention. Figure 3 As shown, the method comprises the following steps:

[0093] Step S301: Obtaining user basic information data and user behavior data, wherein the user basic information data includes age, education, social security record information and information on applied credit lines, and the user behavior data includes the number of times the user clicks on credit products, the number of times the webpage where the credit products are located is browsed, and the user searches for keywords for credit products;

[0094] Step S302: generating a user credit qualification value based on the user basic information data, and reviewing the user's credit qualification based on the user credit qualification value;

[0095] Step S303: Obtaining basic information data and user behavior data of users with credit qualifications, and constructing user portraits based on the basic information data and user behavior data to obtain user credit intention characteristics;

[0096] Step S304: matching credit products based on the user's credit intention characteristics and recommending the credit products to the user.

[0097] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0098] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0099] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0100] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0101] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0102] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A personalized credit product recommendation system based on user portraits, characterized by: The system includes a user data acquisition module, a user credit qualification review module, a user portrait construction module and a credit product matching module; The user data acquisition module is used to acquire user basic information data and user behavior data, wherein the user basic information data includes age, education, social security record information and information on applied credit lines, and the user behavior data includes the number of times the user clicks on credit products, the number of times the webpage where the credit products are located is browsed, and the keywords that the user searches for credit products; The user credit qualification review module is used to generate a user credit qualification value based on the user's basic information data, and review the user's credit qualification based on the user's credit qualification value; The user profile building module is used to obtain basic information data and user behavior data of users with credit qualifications, and to build user profiles based on the basic information data and user behavior data to obtain user credit intention characteristics; The credit product matching module is used to match credit products based on user credit intention characteristics and recommend credit products to users.

2. The personalized credit product recommendation system based on user portrait according to claim 1, characterized in that: Generating a user's credit qualification value based on the user's basic information data includes the following processes: Obtain the user's age A based on the user's basic information data; Based on the user's basic information data, the user's education level is obtained, and the user's corresponding credit qualification basic value B is queried based on the user's education level. Among them, the credit qualification basic value corresponding to a high school education level is 10, the credit qualification basic value corresponding to a bachelor's degree is 20, and the credit qualification basic value corresponding to a postgraduate degree is 30; Based on the user's basic information data, obtain the user's social security record information, and based on the social security record information, obtain the number of consecutive social security payment months C of the user; Based on the basic information data of the user, the information on the credit line applied for is obtained, and based on the information on the credit line applied for, the total amount of credit applied for by the user D is obtained; Substitute age A, credit qualification base value B, consecutive social security payment months C, and user application credit amount D into the user credit qualification value calculation formula to obtain the user credit qualification value, which is as follows: Among them, LMS is the user's credit qualification value.

3. The personalized credit product recommendation system based on user portrait according to claim 1, characterized in that: The credit qualification review of a user based on the user's credit qualification value includes the following process: A credit qualification value threshold is obtained, wherein the credit qualification value threshold is stored in the system and is set by the system to determine whether the user's credit qualification value exceeds the credit qualification value threshold. If so, the user is credit qualified, and if not, the user is not credit qualified.

4. The personalized credit product recommendation system based on user portrait according to claim 1, characterized in that: Build user portraits based on user basic information data and user behavior data to obtain user credit intention characteristics The process includes: The user's basic information data and user behavior data are recorded as text, and useless characters, segmentation words and stop words are removed from the text. The text is segmented using the Jieba Chinese segmentation component to obtain a word set of the text, where the words include age, education, social security, personal tax payment, click, browse and search; Train the word set of the text based on the pre-trained word2vec model, and then use the trained model to generate word vectors; The word vectors of the text are weighted averaged through the TF-IDF weighted statistical method to generate an average word vector to represent the text features. The final text feature vector is TF = {t1, t2...t k }, TF={t1,t2...t k } is recorded as the user's credit intention feature.

5. The personalized credit product recommendation system based on user portrait according to claim 4, characterized in that: The word vectors of the text are weighted averaged through the TF-IDF weighted statistical method to generate an average word vector to represent the text features. The process includes: TF-IDF=TF w ×IDF; The weights are obtained based on the TF-IDF value. Common words are given smaller weights, while less common words are given larger weights. That is, key feature words are extracted through weighted statistics. Among them, f(w) represents the number of times word w appears in the text, TF w is the word frequency, It indicates the number of times other words appear in the text, n is the number of types of other words, and the value of i is 1, 2...n; dfw represents the number of texts in which word w appears.

6. The personalized credit product recommendation system based on user portrait according to claim 1, characterized in that: Matching credit products based on user credit intention characteristics and recommending credit products to users includes the following processes: Use a neural network model to divide the dataset into training and testing sets; Use the training set data to train the model and adjust the model parameters through cross-validation and other methods; Monitor the loss function during training and the performance metrics on the validation set to avoid overfitting; Use the test set to evaluate the performance of the model. Common evaluation indicators include accuracy, recall, F1 score, and AUC-ROC. Tune the model based on the evaluation results, including adjusting model parameters and trying different model structures or algorithms; Predict users’ credit intentions and preferences based on the trained model; Based on the prediction results, the credit products that best meet the needs of users are recommended to them. The personalized recommendation algorithm includes a collaborative filtering algorithm, which combines the user's historical behavior and the behavior of similar users to recommend products.

7. A personalized credit product recommendation method based on user portraits, characterized in that: A personalized credit product recommendation system based on user portrait applicable to any one of claims 1 to 6, the method comprising: Obtaining basic user information and behavior data, where basic user information includes age, education, social security record information, and information on credit lines applied for; and user behavior data includes the number of times a user clicks on a credit product, the number of times a user browses the webpage where the credit product is located, and the keywords that the user searches for the credit product; Generate a user credit qualification value based on the user's basic information data, and review the user's credit qualification based on the user's credit qualification value; Obtain basic information data and user behavior data of users with credit qualifications, and construct user portraits based on the basic information data and user behavior data to obtain user credit intention characteristics; Match credit products based on user credit intention characteristics and recommend credit products to users.

Images