Credit card recommendation method and device, electronic equipment and storage medium

Abstract

The application discloses a credit card recommendation method and device, electronic equipment and a storage medium. On the one hand, the credit card recommendation method provided by the application can determine the appeal of a target user according to the comments of the target user, and then determine the target credit card to be recommended according to the appeal of the target user, so that the credit card recommendation is more targeted, the user pain points are matched, the accuracy of the credit card recommendation is improved, and the credit card recommendation efficiency is improved. On the other hand, when determining the appeal of the target user, the positive emotional comments are filtered out according to the emotional information corresponding to the comments, which not only reduces the number of comments and the calculation amount and calculation time required for credit card recommendation, but also retains the target comments containing user appeal information, thereby improving the accuracy of the credit card recommendation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of credit card recommendation, and particularly relates to a credit card recommendation method and device, electronic equipment and a storage medium. BACKGROUND

[0002] A credit card is a credit certificate issued by a credit card company to a consumer with credit qualification, and can be used for advance consumption in the form of overdraft. In order to match the needs of different consumers, the credit card company designs various types of credit cards, and consumers can select different types of credit cards according to their own needs. However, as the types of credit cards designed by the credit card company increase, there are more and more credit cards with similar benefits. In addition to the benefits of the credit card, the consumer also needs to understand the evaluation of each type of credit card in detail, so as to select the ideal credit card.

[0003] The current credit card recommendation method needs the user to search in the search bar based on his own needs to obtain the ideal credit card type. However, this method usually recommends according to the heat ranking, and the accuracy of the recommended credit card is not high. SUMMARY

[0004] The present application provides a credit card recommendation method, device, electronic equipment and storage medium, which aims to solve the problem that the current credit card recommendation method usually recommends according to the heat ranking, and the accuracy of the recommended credit card is not high.

[0005] In a first aspect, the present application provides a credit card recommendation method, comprising:

[0006] receiving a credit card application instruction, determining a target user corresponding to the credit card application instruction;

[0007] querying the historical comments of the target user from a preset comment database;

[0008] performing prediction processing on the historical comments through a preset sentiment prediction model to obtain sentiment information corresponding to the historical comments, wherein the sentiment information includes one of positive information, neutral information and negative information;

[0009]

[0009] filtering the historical comments according to the sentiment information corresponding to the historical comments to obtain target comments, wherein the sentiment information corresponding to the target comments is one of neutral information and negative information;

[0010] extracting user demand information in the target comments;

[0011] matching the user demand information with benefit information corresponding to a preset candidate credit card to obtain a first similarity between the user demand information and the benefit information;

[0012] According to the first similarity between the user appeal information and the benefit information, target benefit information and a target credit card corresponding to the target benefit information are determined, and the target credit card is recommended to a source terminal of the credit card application instruction.

[0013] In a possible implementation of the present application, the determining, according to the first similarity between the user appeal information and the benefit information, of target benefit information and a target credit card corresponding to the target benefit information, and the recommending of the target credit card to a target terminal of the target user, comprises:

[0014] obtaining user attribute information of the target user, wherein the user attribute information comprises at least one of age, gender and income;

[0015] correcting an initial similarity threshold according to a correction value corresponding to the user attribute information to obtain a preset first similarity threshold;

[0016] comparing the first similarity between the user appeal information and the benefit information with the first similarity threshold to obtain target benefit information with a similarity greater than the first similarity threshold;

[0017] taking a candidate credit card corresponding to the target benefit information as the target credit card, and recommending the target credit card to a target terminal of the target user.

[0018] In a possible implementation of the present application, before the matching of the user appeal information with benefit information corresponding to a preset candidate credit card to obtain the first similarity between the user appeal information and the benefit information, the method further comprises:

[0019] obtaining an initial credit card and benefit information corresponding to the initial credit card;

[0020] classifying the initial credit card according to a benefit type in the benefit information to obtain a plurality of credit card sets;

[0021] obtaining a credit card with the highest score in each credit card set, and taking the obtained credit card as a candidate credit card.

[0022] In a possible implementation of the present application, the extracting of the user appeal information in the target comment comprises:

[0023] performing word segmentation processing on the target comment to obtain candidate words corresponding to the target comment;

[0024] counting the number of occurrences of the candidate words in the target comment to obtain target words with a number of occurrences greater than a preset number threshold.

[0025] The information of the target word is set as user appeal information in the target comment.

[0026] In a possible implementation of the present application, the historical comment of the target user is obtained from the preset comment database, including:

[0027] The user comment of the target user in the preset comment database is obtained;

[0028] The user comment is compared with a preset default comment to obtain a second similarity between the user comment and the default comment;

[0029] The user comment with a second similarity less than a preset second similarity threshold is set as the historical comment of the target user.

[0030] In a possible implementation of the present application, before the historical comment of the target user is obtained from the preset comment database, the method further includes:

[0031] The historical recommendation times corresponding to the target user and the historical recommended credit card corresponding to the target user are obtained;

[0032] If the historical recommendation times are less than a preset times threshold, or the historical recommended credit card is contained in the credit card held by the target user, the step of obtaining the historical comment of the target user from the preset comment database is executed.

[0033] In a possible implementation of the present application, before the historical comment is processed by the preset sentiment prediction model to obtain the sentiment information corresponding to the historical comment, the method further includes:

[0034] A sample data set is obtained, wherein the sample data set includes sample comments and label information corresponding to the sample comments;

[0035] The sample comments are processed by an initial sentiment prediction model to obtain sentiment information corresponding to the sample comments;

[0036] According to the label information and the sentiment information, parameters in the initial sentiment prediction model are adjusted to obtain a preset sentiment prediction model.

[0037] In a second aspect, the present application provides a credit card recommendation device, including:

[0038] A receiving unit is configured to receive a credit card application instruction and determine a target user corresponding to the credit card application instruction;

[0039] The query unit is configured to query historical comments of the target user from a preset comment database;

[0040] The prediction unit is configured to perform prediction processing on the historical comments by using a preset sentiment prediction model to obtain sentiment information corresponding to the historical comments, wherein the sentiment information comprises one of positive information, neutral information, and negative information.

[0041] The screening unit is configured to screen the historical comments according to the sentiment information corresponding to the historical comments to obtain target comments, wherein the sentiment information corresponding to the target comments is one of neutral information and negative information.

[0042] The extraction unit is configured to extract user appeal information in the target comments.

[0043] The matching unit is configured to match the user appeal information with benefit information corresponding to a preset candidate credit card to obtain a first similarity between the user appeal information and the benefit information.

[0044] The determination unit is configured to determine target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommend the target credit card to a source terminal of the credit card application instruction.

[0045] In a possible implementation of the present application, the determination unit is further configured to:

[0046] Obtain user attribute information of the target user, wherein the user attribute information comprises at least one of age, gender, and income.

[0047] Correct an initial similarity threshold according to a correction value corresponding to the user attribute information to obtain a preset first similarity threshold.

[0048] Compare the first similarity between the user appeal information and the benefit information with the first similarity threshold to obtain target benefit information with a similarity greater than the first similarity threshold.

[0049] Recommend the target credit card corresponding to the target benefit information as a target credit card to a target terminal of the target user.

[0050] In a possible implementation of the present application, the matching unit is further configured to:

[0051] Obtain an initial credit card and benefit information corresponding to the initial credit card.

[0052] According to the equity type in the equity information, the initial credit card is classified to obtain a plurality of credit card sets;

[0053] A credit card with the highest score in each credit card set is obtained as a candidate credit card.

[0054] In a possible implementation of the present application, the extraction unit is further configured to:

[0055] The target review is subjected to word segmentation processing to obtain candidate words corresponding to the target review;

[0056] The number of occurrences of the candidate words in the target review is counted to obtain a target word with a number of occurrences greater than a preset number threshold;

[0057] The information of the target word is set as user appeal information in the target review.

[0058] In a possible implementation of the present application, the query unit is further configured to:

[0059] The user review of the target user in a preset review database is obtained;

[0060] The user review is compared with a preset default review to obtain a second similarity between the user review and the default review;

[0061] The user review with a second similarity less than a preset second similarity threshold is set as a historical review of the target user.

[0062] In a possible implementation of the present application, the query unit is further configured to:

[0063] The historical recommendation number corresponding to the target user and the historical recommended credit card corresponding to the target user are obtained;

[0064] If the historical recommendation number is less than a preset number threshold, or the historical recommended credit card is contained in the credit card held by the target user, the step of querying the historical review of the target user from the preset review database is performed.

[0065] In a possible implementation of the present application, the prediction unit is further configured to:

[0066] A sample data set is obtained, wherein the sample data set includes sample reviews and label information corresponding to the sample reviews;

[0067] The sample reviews are subjected to prediction processing by an initial sentiment prediction model to obtain sentiment information corresponding to the sample reviews;

[0068] According to the label information and the sentiment information, parameters in the initial sentiment prediction model are adjusted to obtain a preset sentiment prediction model.

[0069] In a third aspect, the present application also provides an electronic device, which comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, and the processor executes the steps of any credit card recommendation method provided by the present application when invoking the computer program in the memory.

[0070] In a fourth aspect, the present application also provides a storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of any credit card recommendation method provided by the present application.

[0071] To sum up, the credit card recommendation method provided by the embodiments of the present application comprises: receiving a credit card application instruction, determining a target user corresponding to the credit card application instruction; querying a historical comment of the target user from a preset comment database; performing prediction processing on the historical comment by a preset sentiment prediction model to obtain sentiment information corresponding to the historical comment, wherein the sentiment information comprises one of positive information, neutral information and negative information; filtering the historical comment according to the sentiment information corresponding to the historical comment to obtain a target comment, wherein the sentiment information corresponding to the target comment is one of neutral information and negative information; extracting user appeal information in the target comment; matching the user appeal information with benefit information corresponding to a preset candidate credit card to obtain a first similarity between the user appeal information and the benefit information; determining target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommending the target credit card to a source terminal of the credit card application instruction.

[0072] On the one hand, the credit card recommendation method provided by the embodiments of the present application can determine the appeal of the target user according to the comment of the target user, and then determine the target credit card to be recommended according to the appeal of the target user, so that the credit card recommendation is more targeted, the user pain points are matched, the accuracy of the credit card recommendation is improved, and the credit card recommendation is automatically performed, thereby improving the efficiency of the credit card recommendation. On the other hand, when determining the appeal of the target user, the comments with positive sentiment are filtered out according to the sentiment information corresponding to the comments, which not only reduces the number of comments and the amount of calculation and the time of calculation required when recommending the credit card, but also retains the target comments containing the user appeal information, thereby improving the accuracy of the credit card recommendation. BRIEF DESCRIPTION OF DRAWINGS

[0073] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments description. Obviously, the drawings in the following description only constitute some embodiments of the present application. For those skilled in the art, other drawings can also be obtained from these drawings without any creative effort.

[0074] Figure 1 is a schematic diagram of an application scenario of the credit card recommendation method provided by the embodiments of the present application;

[0075] Figure 2 is a schematic diagram of a flow of the credit card recommendation method provided by the embodiments of the present application;

[0076] Figure 3 is a schematic diagram of a flow of the credit card recommendation method provided by the embodiments of the present application;

[0077] Figure 4 is a schematic diagram of another flow of the credit card recommendation method provided by the embodiments of the present application;

[0078] Figure 5 is a schematic diagram of an embodiment structure of the credit card recommendation device provided by the embodiments of the present application;

[0079] Figure 6 is a schematic diagram of an embodiment structure of the electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0080] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort fall within the scope of protection of the present application.

[0081] In the description of the embodiments of the present application, it should be understood that the terms "first", "second" are only used for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second" can explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0082] The following description is presented to enable any person skilled in the art to practice the application as claimed. In the following description, for purposes of explanation, specific details are set forth to provide a thorough understanding of the application. It is apparent, however, to one skilled in the art that the application can be practiced without using these specific details. In other instances, well-known processes have not been elaborated as not to obscure the description of the present embodiments of the application. Thus, the present application is not intended to be limited by the embodiments shown, but is to be accorded with the widest scope consistent with the principles and features disclosed herein.

[0083] Embodiments of the present application provide a credit card recommendation method, device, electronic device and storage medium. The credit card recommendation device can be integrated in an electronic device, which can be a server or a terminal device.

[0084] The credit card recommendation method can be executed by the credit card recommendation device provided by the embodiments of the present application, or a server device, a physical host or a user equipment (UE) and other electronic devices integrated with the credit card recommendation device. The credit card recommendation device can be implemented in hardware or software. The UE can be a terminal device such as a smart phone, a tablet computer, a notebook computer, a palm computer, a desktop computer or a personal digital assistant (PDA).

[0085] The electronic device can be operated independently or in a device cluster.

[0086] Referring to Figure 1 , Figure 1 is a scene diagram of the credit card recommendation system provided by the embodiments of the present application. The credit card recommendation system can include an electronic device 101, and the credit card recommendation device is integrated in the electronic device 101.

[0087] In addition, as Figure 1 shown, the credit card recommendation system can further include a memory 102 for storing data, such as text data.

[0088] It should be noted that Figure 1 the scene diagram of the credit card recommendation system shown is only an example. The credit card recommendation system and the scene described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not limit the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, as the credit card recommendation system evolves and new business scenarios emerge, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0089] Next, the credit card recommendation method provided by the embodiments of the present application will be introduced. In the embodiments of the present application, an electronic device is taken as an execution subject. In order to simplify and facilitate the description, the execution subject will be omitted in the following method embodiments. The credit card recommendation method comprises the following steps: receiving a credit card application instruction, determining a target user corresponding to the credit card application instruction; querying historical comments of the target user from a preset comment database; performing prediction processing on the historical comments by using a preset sentiment prediction model to obtain sentiment information corresponding to the historical comments, wherein the sentiment information comprises one of positive information, neutral information and negative information; screening the historical comments according to the sentiment information corresponding to the historical comments to obtain target comments, wherein the sentiment information corresponding to the target comments is one of neutral information and negative information; extracting user appeal information in the target comments; matching the user appeal information with benefit information corresponding to a preset candidate credit card to obtain a first similarity between the user appeal information and the benefit information; determining target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommending the target credit card to a source terminal of the credit card application instruction.

[0090] Reference Figure 2 , Figure 2 is a flowchart of the credit card recommendation method provided by the embodiments of the present application. It should be noted that although a logical sequence is shown in the flowchart, in some cases, the steps shown or described herein can be performed in an order different from that shown herein. The credit card recommendation method can specifically comprise the following steps 201-205, wherein:

[0091] 201. Receive a credit card application instruction, and determine a target user corresponding to the credit card application instruction.

[0092] The credit card recommendation method provided by the embodiments of the present application can be applied to the financial field. The credit card application instruction can be an instruction issued by a bank software when a user applies for a new credit card. For example, when a user clicks a virtual button for applying for a new credit card such as “apply for a card” on a bank software operated by a terminal such as a smart phone or a personal computer, it is equivalent to issuing a credit card application instruction.

[0093] The user identifier carried by the credit card application instruction can be used by the electronic device to obtain the corresponding target user. For example, the user identifier can be a terminal device number of the terminal operated by the user at this time, or can be an account name used by the user, and the like.

[0094] 202. Query historical comments of the target user from a preset comment database.

[0095] The preset comment database can refer to a database for storing comments. In the embodiments of the present application, the comments can refer to comments of a user on a credit card. For example, a database for storing comments in a bank software background can be used as the preset comment database.

[0096] In step 202, the electronic device can query the historical comments of the target user from the comment database according to the user identifier corresponding to the target user.

[0097] In some embodiments, the default comment sent by the software after the target user opens a card without timely commenting can also be deleted by a certain method, so as to reduce the number of historical comments and retain real comments with use value. At this time, the step of querying the historical comments of the target user from the preset comment database includes:

[0098] (1.1) Obtaining the user comments of the target user in the preset comment database.

[0099] The preset comment database is described above and will not be described in detail.

[0100] The user comments of the target user can refer to all comments published by the target user in the preset comment database.

[0101] (1.2) Comparing the user comments with the preset default comments to obtain a second similarity between the user comments and the default comments.

[0102] The preset default comment refers to a comment sent by the software after the user opens a card without timely commenting.

[0103] When comparing, the electronic device can convert the user comments and the preset default comments into word vectors through an open-source language processing model such as word2vec, and then compare the similarity between the two word vectors to obtain the second similarity.

[0104] It can be understood that the greater the second similarity, the more similar the user comments and the default comments, and the greater the probability that the user comments are comments sent by the software after the user opens a card without timely commenting.

[0105] (1.3) Setting the user comments with a second similarity less than a preset second similarity threshold as the historical comments of the target user.

[0106] Through the method of step (1.1) to step (1.3), the default comments can be effectively removed, and the comments with use value can be retained, so as to reduce the subsequent calculation amount.

[0107] 203、through the preset emotion prediction model, the historical comments are predicted to obtain the emotion information corresponding to the historical comments, wherein the emotion information includes one of positive information, neutral information and negative information.

[0108] The preset emotion prediction model can be used to predict the emotion information in the text. For example, an open source model such as ABSA (Aspect Based Sentiment Analysis) can be used as an initial emotion prediction model. After training the initial emotion prediction model, the preset emotion prediction model is obtained. The emotion information in the historical comments is predicted by the preset emotion prediction model to obtain the attitude of the target user to different credit cards. In the embodiments of the present application, the emotion information includes one of positive information, neutral information and negative information. If the output emotion information is positive information, it means that the target user has a positive attitude towards the credit card corresponding to the historical comments, and the credit card meets the use demand of the target user. If the output emotion information is neutral information or negative information, it means that the target user has a neutral attitude or a negative attitude towards the credit card corresponding to the historical comments, and the credit card does not completely meet the use demand of the target user. For example, when the emotion information is positive information such as like and praise, it means that the target user has a positive attitude towards the credit card corresponding to the historical comments, and the credit card meets the use demand of the target user. When the emotion information is neutral information or negative information such as cold and criticism, it means that the target user has a neutral attitude or a negative attitude towards the credit card corresponding to the historical comments, and the credit card does not completely meet the use demand of the target user.

[0109] The initial emotion prediction model can be trained by the following method:

[0110] (2.1) Obtain a sample data set, wherein the sample data set includes sample comments and label information corresponding to the sample comments.

[0111] The sample comments can also be extracted from the preset comment database. The label information corresponding to the sample comments can be obtained by manual annotation. It can be understood that the label information is also one of positive information, neutral information and negative information.

[0112] (2.2) The sample comments are predicted by the initial emotion prediction model to obtain the emotion information corresponding to the sample comments.

[0113] (2.3) According to the label information and the emotion information, the parameters in the initial emotion prediction model are adjusted to obtain the preset emotion prediction model.

[0114] 204、screen the historical comments according to the corresponding sentiment information of the historical comments, to obtain target comments, wherein the corresponding sentiment information of the target comments is one of neutral information and negative information.

[0115] In some embodiments, the comments in the historical comments corresponding to the positive information can be excluded, and the remaining historical comments can be used as the target comments.

[0116] The purpose of screening is to obtain comments published when the user is not completely satisfied, and then the reasons why the user is not satisfied with the credit card can be obtained from these comments, and suitable credit cards can be recommended according to these reasons.

[0117] 205、extract user demand information in the target comments.

[0118] In some embodiments, the electronic device can obtain the user demand information in the target comments through a preset semantic recognition model. Wherein, an open source text classification network can be used as an initial semantic recognition model, and then the initial semantic recognition model is trained through sample data to obtain the preset semantic recognition model. It can be understood that at this time, the user demand information extracted from a historical comment is one of the preset demand information.

[0119] In other embodiments, the user demand information in the target comments can be extracted according to the number of occurrences of words in the user demand information. At this time, the step of "extracting user demand information in the target comments" includes:

[0120] (3.1) performing word segmentation processing on the target comments to obtain candidate words corresponding to the target comments.

[0121] Exemplarily, the electronic device can perform word segmentation processing on the target comments through an open source language processing model such as word2vec to obtain candidate words corresponding to the target comments. For example, when the target comment is "The repayment interest of this credit card is particularly high, and the interest is as high as XX%", the candidate words obtained after word segmentation include "this", "credit card", "repayment", "interest", "particularly", "surprisingly", and "XX%".

[0122] In some embodiments, the electronic device can also screen the words obtained after word segmentation, retain the nouns therein, and use the retained nouns as candidate words. For example, the above-mentioned words obtained after word segmentation can be screened to retain the nouns "credit card" and "interest", and "credit card" and "interest" can be used as candidate words.

[0123] (3.2) count the number of occurrences of the candidate words in the target comments to obtain target words with a number of occurrences greater than a preset number threshold.

[0124] The preset number threshold is used to evaluate the size of the number of occurrences of the candidate word, and the specific value can be set according to the actual scene. For example, the preset number threshold can be set to 1, that is, if the candidate word occurs at least twice, it is taken as the target word.

[0125] For example, in the example of step (3.1), the candidate word "interest" occurs twice, and the candidate word "credit card" occurs once, so the candidate word "interest" is taken as the target word.

[0126] The purpose of obtaining the target word is to obtain the reason for the user to post the comment. The more the number of occurrences of the candidate word, the higher the probability that the word is related to the reason for posting the comment. Therefore, the candidate word with more occurrences is taken as the target word.

[0127] (3.3) Set the information of the target word as the user demand information in the target comment.

[0128] After obtaining the target word, the information of the target word can be taken as the user demand information, that is, the target word is taken as the user's demand. For example, in the above example, interest can be taken as the user's demand, that is, the user can demand a credit card with low interest.

[0129] 206, match the user demand information with the preset benefit information corresponding to the candidate credit card to obtain a first similarity between the user demand information and the benefit information.

[0130] In order to obtain a credit card that meets the user's demand, the user demand information can be matched with the benefit information corresponding to the candidate credit card.

[0131] The preset candidate credit card can include all credit cards issued by the bank.

[0132] The benefit information can refer to the user benefit set by the bank for the credit card. For example, it can include N interest-free periods, exclusive customer service, N times of overdue, etc. When the bank designs the credit card, the corresponding benefit information can be associated with the credit card in the background database.

[0133] When step 206 is performed, the electronic device can convert the user demand information and the benefit information corresponding to the candidate credit card into word vectors through a word2vec language processing model, and then compare the two word vectors to obtain the first similarity between the user demand information and the benefit information.

[0134] 207. determine target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommend the target credit card to a terminal from which the credit card application instruction is sent.

[0135] In some embodiments, the first similarity can be compared with a preset first similarity threshold value to obtain a target credit card with a first similarity greater than the first similarity threshold value, and the target credit card can be pushed to the terminal from which the credit card application instruction is sent. For example, after a target user operates a bank software through a terminal such as a smart phone or a personal computer and clicks a virtual button for applying for a new credit card such as an "application for opening card", an electronic device obtains a target credit card through the method of steps 201-207, and then displays the target credit card on an interface of the bank software.

[0136] It should be noted that if there are multiple candidate credit cards with a first similarity greater than the first similarity threshold value, the one with the largest first similarity can be selected as the target credit card, and the target credit card can be pushed to the terminal.

[0137] It can be seen that the target credit card obtained through the method of steps 201-207 has benefit information corresponding to the credit card that matches the appeal of the user, and thus meets the needs of the user.

[0138] In some embodiments, the first similarity threshold value can also be adaptively adjusted according to the user attributes of the target user. At this time, the step "determine target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommend the target credit card to a terminal from which the credit card application instruction is sent" includes:

[0139] (4.1) obtain user attribute information of the target user, wherein the user attribute information includes at least one of age, gender, and income.

[0140] When step (4.1) is performed, an electronic device can read the user attribute information of the target user from a preset user attribute database according to a user identifier of the target user.

[0141] The preset user attribute database can be a database used by a bank software background to store user attribute information.

[0142] (4.2) correct an initial similarity threshold value according to a correction value corresponding to the user attribute information to obtain a preset first similarity threshold value.

[0143] The initial similarity threshold value is a preset reference threshold value, and the specific value can be set according to actual scene requirements.

[0144] The correction value is a value for adjusting the initial similarity threshold value, and different correction values can be used to adjust the initial similarity threshold value when the user attribute information is different. For example, for an older user, the benefits he or she wishes to enjoy can be relatively fixed, and even if the benefits corresponding to the credit card are relatively close to the benefits he or she wishes to enjoy, the user can not apply for the credit card. For example, for an older user, if the benefits he or she wishes to enjoy are related to "low interest" or "interest-free", the user can not apply for a credit card with benefits of "interest can be phased". For a younger user, as long as the benefits corresponding to the credit card are associated with the benefits he or she wishes to enjoy, the user can apply for the credit card.

[0145] When step (4.2) is performed, the electronic device can query the preset correspondence table to obtain the correction value corresponding to the user attribute information. Then, the difference between the initial similarity threshold value and the correction value is calculated to obtain the preset first similarity threshold value.

[0146] The preset correspondence table can be stored in the background database of the bank software.

[0147] (4.3) Comparing the first similarity between the user appeal information and the benefit information with the first similarity threshold value to obtain target benefit information with a similarity greater than the first similarity threshold value.

[0148] (4.4) Taking the candidate credit card corresponding to the target benefit information as a target credit card, and recommending the target credit card to a target terminal of a target user.

[0149] In summary, the credit card recommendation method provided by the embodiments of the present application includes: receiving a credit card application instruction, determining a target user corresponding to the credit card application instruction; querying the historical comments of the target user from a preset comment database; performing prediction processing on the historical comments through a preset sentiment prediction model to obtain sentiment information corresponding to the historical comments, wherein the sentiment information includes one of positive information, neutral information and negative information; filtering the historical comments according to the sentiment information corresponding to the historical comments to obtain target comments, wherein the sentiment information corresponding to the target comments is one of neutral information and negative information; extracting user appeal information in the target comments; matching the user appeal information with benefit information corresponding to a preset candidate credit card to obtain a first similarity between the user appeal information and the benefit information; determining target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommending the target credit card to a terminal from which the credit card application instruction is obtained.

[0150] In one aspect, the credit card recommendation method provided by the embodiments of the present application can determine the appeal of a target user according to the comments of the target user, and then determine the target credit card to be recommended according to the appeal of the target user, so that the recommendation of the credit card is more targeted, the user pain points are matched, the accuracy of the credit card recommendation is improved, and the credit card recommendation efficiency is improved. In another aspect, when determining the appeal of the target user, the comments are filtered according to the corresponding emotional information, the comments with positive emotions are filtered out, the number of comments is reduced, the calculation amount and calculation time required when recommending the credit card are reduced, the target comments containing user appeal information are retained, and the accuracy of the credit card recommendation is improved.

[0151] In some embodiments, when selecting a candidate credit card, the credit cards can be classified first, and then the credit cards with higher scores are selected as candidate credit cards to reduce the number of credit cards, and thus the calculation amount and calculation time when obtaining the first similarity can be reduced. Figure 3 At this time, before the step of "matching the user appeal information with the benefit information corresponding to the candidate credit card to obtain the first similarity between the user appeal information and the benefit information", the method further comprises:

[0152] 301. Obtain an initial credit card and benefit information corresponding to the initial credit card.

[0153] In the embodiments of the present application, the initial credit card can include all credit cards issued by a bank.

[0154] 302. Classify the initial credit cards according to the benefit types in the benefit information to obtain a plurality of credit card sets.

[0155] The benefit type can include "interest", "after-sales", "commodity", "financial management", etc. For example, the benefit type corresponding to the benefits related to interest, such as "N period of interest-free", "interest can be divided into installments", etc. can be "interest".

[0156] When performing step 302, the electronic device can classify the initial credit cards with the same benefit type into a category to obtain a credit card set. After classifying all initial credit cards, all credit card sets can be obtained.

[0157] 303. Obtain the credit card with the highest score in each credit card set, and obtain the credit card as a candidate credit card.

[0158] The credit card score can refer to a score stored in a background database of the bank software after the user scores the initial credit card on the bank software. In step 303, for each credit card set, the electronic device can obtain the credit card score corresponding to the credit card from the background database of the bank software, sort the credit card scores corresponding to the credit cards, and take the credit card with the highest credit card score. Then, the credit card taken from each credit card set is taken as a candidate credit card. It can be understood that if there are 10 credit card sets, the number of candidate credit cards is 10.

[0159] The purpose of performing steps 301-303 is to set only one candidate credit card for each benefit type, thereby reducing the number of candidate credit cards, and further reducing the calculation amount and calculation time when obtaining the first similarity. In addition, since there is a corresponding candidate credit card for each benefit type, there is no problem of being unable to determine the target credit card matching the user's appeal information.

[0160] In some embodiments, the electronic device can first determine whether the target user is interested in the recommended credit card, and if the target user is interested in the recommended credit card, perform steps 202-207 to avoid wasting computing resources and causing the target user to be dissatisfied. Referring to Figure 4 At this time, before the step of querying the historical comments of the target user from the preset comment database, the method further includes:

[0161] 401. Obtain the historical recommendation times corresponding to the target user and the historical recommended credit cards corresponding to the target user.

[0162] In the embodiments of the present application, the historical recommendation times refer to the historical recommendation times of the credit card. Therefore, the historical recommendation times corresponding to the target user refer to the times of recommending credit cards to the target user.

[0163] The historical recommended credit card refers to the credit card that has been recommended to the user. Therefore, the historical recommended credit card corresponding to the target user refers to the credit card that has been recommended to the target user.

[0164] The historical recommendation times corresponding to the target user and the historical recommended credit cards corresponding to the target user can be stored in the background database of the bank software, and the electronic device reads the historical recommendation times and the historical recommended credit cards therefrom in step 401.

[0165] 402. If the historical recommendation times are less than a preset number threshold, or the credit card held by the target user contains the historical recommended credit card, perform the step of querying the historical comments of the target user from the preset comment database.

[0166] If the number of historical recommendations is less than the preset number threshold, it indicates that the number of times of recommending the credit card to the target user is less. Even if the target user has not opened the recommended credit card, it does not mean that the target user is not interested in the recommended credit card. Therefore, the electronic device can execute steps 202-207.

[0167] If the credit card held by the target user contains the historical recommended credit card, it indicates that the target user has opened the recommended credit card. Therefore, it can be concluded that the target user is interested in the recommended credit card. The electronic device can execute steps 202-207.

[0168] In order to better implement the credit card recommendation method in the embodiments of the present application, on the basis of the credit card recommendation method, the credit card recommendation device in the embodiments of the present application further provides a credit card recommendation device, as shown in Figure 5 The credit card recommendation device 500 includes:

[0169] The receiving unit 501 is configured to receive a credit card application instruction, and determine a target user corresponding to the credit card application instruction.

[0170] The querying unit 502 is configured to query historical comments of the target user from a preset comment database.

[0171] The prediction unit 503 is configured to perform prediction processing on the historical comments by using a preset sentiment prediction model, to obtain sentiment information corresponding to the historical comments, wherein the sentiment information includes one of positive information, neutral information, and negative information.

[0172] The screening unit 504 is configured to screen the historical comments according to the sentiment information corresponding to the historical comments, to obtain target comments, wherein the sentiment information corresponding to the target comments is one of neutral information and negative information.

[0173] The extraction unit 505 is configured to extract user appeal information in the target comments.

[0174] The matching unit 506 is configured to match the user appeal information with benefit information corresponding to a preset candidate credit card, to obtain a first similarity between the user appeal information and the benefit information.

[0175] The determination unit 507 is configured to determine target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommend the target credit card to a source terminal of the credit card application instruction.

[0176] In a possible implementation manner of the present application, the determination unit 507 is further configured to:

[0177] obtaining user attribute information of the target user, wherein the user attribute information comprises at least one of age, gender and income;

[0178] correcting an initial similarity threshold according to a correction value corresponding to the user attribute information, to obtain a preset first similarity threshold;

[0179] comparing a first similarity between the user appeal information and the benefit information with the first similarity threshold, to obtain target benefit information with a similarity greater than the first similarity threshold;

[0180] taking a candidate credit card corresponding to the target benefit information as a target credit card, and recommending the target credit card to a target terminal of the target user.

[0181] In a possible implementation manner of the present application, the matching unit 506 is further configured to:

[0182] obtain an initial credit card and benefit information corresponding to the initial credit card;

[0183] classify the initial credit card according to a benefit type in the benefit information, to obtain a plurality of credit card sets;

[0184] obtain a credit card with the highest score in each credit card set, and take the obtained credit card as a candidate credit card.

[0185] In a possible implementation manner of the present application, the extraction unit 505 is further configured to:

[0186] perform word segmentation processing on the target comment, to obtain candidate words corresponding to the target comment;

[0187] count the number of occurrences of the candidate words in the target comment, to obtain a target word with a number of occurrences greater than a preset number threshold;

[0188] set the information of the target word as user appeal information in the target comment.

[0189] In a possible implementation manner of the present application, the query unit 502 is further configured to:

[0190] obtain a user comment of the target user in a preset comment database;

[0191] compare the user comment with a preset default comment, to obtain a second similarity between the user comment and the default comment;

[0192] set a user comment with a second similarity less than a preset second similarity threshold as a historical comment of the target user.

[0193] In a possible implementation of the present application, the query unit 502 is further configured to:

[0194] obtain a historical recommendation number of times corresponding to the target user and a historical recommended credit card corresponding to the target user;

[0195] If the historical recommendation number of times is less than a preset number threshold, or the historical recommended credit card is included in the credit cards held by the target user, the step of querying the historical comment of the target user from the preset comment database is performed.

[0196] In a possible implementation of the present application, the prediction unit 503 is further configured to:

[0197] obtain a sample data set, wherein the sample data set includes a sample comment and label information corresponding to the sample comment;

[0198] perform prediction processing on the sample comment by using an initial sentiment prediction model to obtain sentiment information corresponding to the sample comment;

[0199] adjust parameters in the initial sentiment prediction model according to the label information and the sentiment information to obtain a preset sentiment prediction model.

[0200] In implementation, each of the above units can be implemented as an independent entity, or can be combined as the same or several entities, and the specific implementation of each of the above units can be referred to the method embodiments above, which will not be described here.

[0201] Since the credit card recommendation apparatus can perform the steps in the credit card recommendation method in any embodiment, the beneficial effects of the credit card recommendation method in any embodiment of the present application can be achieved, which are described in detail above and will not be described here.

[0202] In addition, in order to better implement the credit card recommendation method in the embodiments of the present application, based on the credit card recommendation method, the embodiments of the present application further provide an electronic device, which is described in detail below. Figure 6 , Figure 6 A structural schematic diagram of an electronic device according to an embodiment of the present application is shown, and specifically, the electronic device provided by the embodiments of the present application includes a processor 601, which is configured to implement each step of the credit card recommendation method in any embodiment when executing a computer program stored in a memory 602; or the processor 601 is configured to implement the functions of each module in the corresponding embodiment when executing the computer program stored in the memory 602. Figure 5

[0203] ​For example, the computer program can be divided into one or more modules / units, one or more modules / units are stored in the memory 602 and executed by the processor 601 to complete the embodiments of the present application. One or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the computer device.

[0204] The electronic device can include, but is not limited to, the processor 601, the memory 602. Those skilled in the art can understand that the schematic diagram is only an example of the electronic device and does not constitute a limitation on the electronic device, which can include more or fewer components than the schematic diagram, or combine certain components, or different components.

[0205] The processor 601 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor and the like. The processor is the control center of the electronic device, which connects all parts of the electronic device through various interfaces and lines.

[0206] The memory 602 can be used to store computer programs and / or modules. The processor 601 realizes various functions of the computer device by running or executing the computer programs and / or modules stored in the memory 602, and calling the data stored in the memory 602. The memory 602 can mainly include a program storage area and a data storage area, wherein the program storage area can store operating systems, application programs required by at least one function (such as sound playing function, image playing function, etc.), etc.; the data storage area can store data created according to the use of the electronic device (such as audio data, video data, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state memory device.

[0207] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the credit card recommendation device, the electronic device and the corresponding units described above can refer to the description of the credit card recommendation method in any embodiment, and will not be repeated here.

[0208] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by controlling related hardware by instructions, which can be stored in a storage medium and loaded and executed by a processor.

[0209] To this end, an embodiment of the present application provides a storage medium, and the storage medium stores a computer program. The computer program is executed by a processor to perform the steps of the credit card recommendation method in any embodiment of the present application. For specific operations, refer to the description of the credit card recommendation method in any embodiment, which will not be repeated here.

[0210] The storage medium can include a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0211] Since the instructions stored in the storage medium can perform the steps of the credit card recommendation method in any embodiment of the present application, the beneficial effects of the credit card recommendation method in any embodiment of the present application can be achieved. For details, refer to the previous description, which will not be repeated here.

[0212] The above provides a credit card recommendation method, device, storage medium and electronic device. The principle and implementation mode of the present application are described by applying specific examples. The above embodiment is only used to help understand the method and its core idea of the present application. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A credit card recommendation method characterized by comprising: The method comprises the following steps: receiving a credit card application instruction, determining a target user corresponding to the credit card application instruction; querying historical comments of the target user from a preset comment database; performing prediction processing on the historical comments by using a preset sentiment prediction model to obtain sentiment information corresponding to the historical comments, wherein the sentiment information comprises one of positive information, neutral information and negative information; screening the historical comments according to the sentiment information corresponding to the historical comments to obtain target comments, wherein the sentiment information corresponding to the target comments is one of neutral information and negative information; extracting user demand information in the target comments; matching the user demand information with benefit information corresponding to a preset candidate credit card to obtain a first similarity between the user demand information and the benefit information; determining target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user demand information and the benefit information, and recommending the target credit card to a source terminal of the credit card application instruction; The method comprises the following steps: obtaining user attribute information of the target user, wherein the user attribute information comprises at least one of age, gender and income; correcting an initial similarity threshold value according to a correction value corresponding to the user attribute information to obtain a preset first similarity threshold value; comparing the first similarity between the user demand information and the benefit information with the first similarity threshold value to obtain target benefit information with a similarity greater than the first similarity threshold value; recommending a candidate credit card corresponding to the target benefit information as a target credit card to a target terminal of the target user.

2. The credit card recommendation method of claim 1, characterized by, Before the matching of the user demand information with the benefit information corresponding to the preset candidate credit card to obtain the first similarity between the user demand information and the benefit information, the method further comprises the following steps: obtaining an initial credit card and benefit information corresponding to the initial credit card; classifying the initial credit card according to a benefit type in the benefit information to obtain a plurality of credit card sets; obtaining a credit card with the highest score in each credit card set, and taking the obtained credit card as a candidate credit card.

3. The credit card recommendation method of claim 1, characterized by, The method comprises the following steps: performing word segmentation processing on the target comments to obtain candidate words corresponding to the target comments; counting the number of occurrences of the candidate words in the target comments to obtain target words with a number of occurrences greater than a preset number threshold value; setting the information of the target words as the user demand information in the target comments.

4. The credit card recommendation method of claim 1, characterized by, The method comprises the following steps: obtaining user comments of the target user in the preset comment database; The user comment is compared with a preset default comment to obtain a second similarity between the user comment and the default comment; The user comment with a second similarity less than a preset second similarity threshold is set as a historical comment of the target user.

5. The credit card recommendation method of claim 1, characterized by, Before the querying the historical comment of the target user from the preset comment database, the method further comprises: obtaining a historical recommendation frequency corresponding to the target user and a historical recommendation credit card corresponding to the target user; if the historical recommendation frequency is less than a preset frequency threshold or the historical recommendation credit card is included in the credit card held by the target user, the step of querying the historical comment of the target user from the preset comment database is executed.

6. The credit card recommendation method according to any one of claims 1 to 5, characterized by, Before the predicting processing of the historical comment by the preset sentiment prediction model to obtain the sentiment information corresponding to the historical comment, the method further comprises: obtaining a sample data set, wherein the sample data set includes sample comments and label information corresponding to the sample comments; predicting processing of the sample comments by an initial sentiment prediction model to obtain sentiment information corresponding to the sample comments; adjusting parameters in the initial sentiment prediction model according to the label information and the sentiment information to obtain a preset sentiment prediction model.

7. A credit card recommendation device characterized by comprising: comprises: a receiving unit configured to receive a credit card application instruction and determine a target user corresponding to the credit card application instruction; a querying unit configured to query a historical comment of the target user from a preset comment database; a predicting unit configured to predict process the historical comment by a preset sentiment prediction model to obtain sentiment information corresponding to the historical comment, wherein the sentiment information includes one of positive information, neutral information and negative information; a screening unit configured to screen the historical comment according to the sentiment information corresponding to the historical comment to obtain a target comment, wherein the sentiment information corresponding to the target comment is one of neutral information and negative information; an extracting unit configured to extract user appeal information in the target comment; a matching unit configured to match the user appeal information with benefit information corresponding to a preset candidate credit card to obtain a first similarity between the user appeal information and the benefit information; a determining unit configured to determine target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommend the target credit card to a source terminal of the credit card application instruction; the determining unit configured to determine target benefit information and a target credit card corresponding to the target benefit information according to the first similarity between the user appeal information and the benefit information, and recommend the target credit card to a source terminal of the credit card application instruction, comprises: obtaining user attribute information of the target user, wherein the user attribute information includes at least one of age, gender and income; correcting an initial similarity threshold according to a correction value corresponding to the user attribute information to obtain a preset first similarity threshold; comparing a first similarity between the user appeal information and the benefit information with the first similarity threshold to obtain target benefit information with a similarity greater than the first similarity threshold; recommending the target credit card to a target terminal of a target user.

8. An electronic device, comprising: The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the steps in the credit card recommendation method of any one of claims 1 to 6.

9. A storage medium, characterized by The storage medium has a computer program stored thereon, and the computer program is executable by a processor to implement the steps in the credit card recommendation method of any one of claims 1 to 6.

Images