Financial product recommendation method, computer equipment and program product

Abstract

The embodiment of the invention provides a financial product recommendation method, which comprises the following steps: receiving behavior information sent by user equipment, the behavior information being behavior information when a user browses a product display interface; determining a target intention of the user according to the behavior information; determining recommended financial product information for the user according to the target intention and multiple pieces of financial product information; and sending the recommended financial product information to the user equipment, wherein the recommended financial product information is used for being displayed on the product display interface. The embodiment of the invention further provides computer equipment and a computer program product.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a financial product recommendation method, computer equipment, and program product. Background Art

[0002] With the development of social economy and the improvement of consumption level, people's consumption of financial products is becoming more and more common. At present, the recommendation of financial products is based on the pre-calculation of the user's existing data by the financial product service platform. For example, the financial product service platform recommends financial products based on the user's historical browsing information, historical ordering information, credit information and other offline information. This kind of pre-calculation lacks real-time performance, cannot meet the current needs of users, and cannot quickly respond to changes in user preferences and behaviors. It is difficult to improve the accuracy and effectiveness of recommendations, which ultimately leads to a low subscription rate for financial product service platforms. Summary of the invention

[0003] Embodiments of the present application provide a financial product recommendation method, a computer device, and a program product.

[0004] The technical solution of this application is implemented as follows:

[0005] In a first aspect, an embodiment of the present application provides a financial product recommendation method, the method comprising:

[0006] Receiving behavior information sent by a user device; the behavior information is behavior information of the user when browsing a product display interface;

[0007] Determining the target intention of the user according to the behavior information;

[0008] Determining recommended financial product information for the user based on the target intention and the plurality of financial product information;

[0009] The recommended financial product information is sent to the user equipment, and the recommended financial product information is used to be displayed on the product display interface.

[0010] In a second aspect, an embodiment of the present application provides a computer device, comprising: a processor and a memory, the memory being used to store a computer program, the processor being used to call and run the computer program stored in the memory, and executing the method described in the first aspect.

[0011] In a third aspect, the present application provides a computer program product, including computer program instructions, which implement the method described in the first aspect when executed by a processor.

[0012] The embodiment of the present application provides a financial product recommendation method, in which a computer device can obtain the user's behavior information sent by the user device for the product display interface, so that the computer device can identify the user's target intention based on the behavior information, and determine the recommended financial product information based on the user's target intention, and feed back the recommended financial product information to the user device; in this way, the user device can display the recommended financial product information on the product display interface. It can be seen that the embodiment of the present application provides an interactive financial product recommendation method, which can determine the user's current intention based on the user's behavior information, and recommend financial products to the user in real time based on the user's intention. In this way, it can quickly respond to the user's preferences and behavior changes, improve the accuracy and effect of the recommendation, and thus increase the subscription rate of financial products. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 A schematic diagram of the architecture of a financial product recommendation system used in an embodiment of the present application;

[0014] Figure 2 A financial product recommendation method provided in this application embodiment Figure 1 ;

[0015] Figure 3 A financial product recommendation method provided in this application embodiment Figure 2 ;

[0016] Figure 4A and Figure 4B A financial product recommendation method provided in this application embodiment Figure 3 ;

[0017] Figure 5 Schematic diagram 4 of a financial product recommendation method provided in an embodiment of the present application;

[0018] Figure 6 A financial product recommendation method provided in this application embodiment Figure 5 ;

[0019] Figure 7 A financial product recommendation method provided in this application embodiment Figure 6 ;

[0020] Figure 8 A flowchart of a reporting method provided in an embodiment of the present application;

[0021] Fig. 9 A schematic diagram of a user intention recognition process provided in an embodiment of the present application;

[0022] Fig.10A schematic diagram of a financial product recall process provided in an embodiment of the present application;

[0023] Fig.11 A schematic diagram of a sorting method flow chart provided in an embodiment of the present application;

[0024] Fig.12 A schematic diagram of an access method flow provided in an embodiment of the present application;

[0025] Fig.13 A schematic diagram of the structure of a financial product recommendation device provided in an embodiment of the present application;

[0026] Fig.14 A schematic diagram of the hardware structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] It should be understood that the term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0029] It should also be understood that the "indication" mentioned in the embodiments of the present application can be a direct indication, an indirect indication, or an indication of an association relationship. For example, A indicates B, which can mean that A directly indicates B, for example, B can be obtained through A; it can also mean that A indirectly indicates B, for example, A indicates C, and B can be obtained through C; it can also mean that there is an association relationship between A and B.

[0030] It should also be understood that the "correspondence" mentioned in the embodiments of the present application may indicate a direct or indirect correspondence between the two, or an association between the two, or a relationship of indication and being indicated, configuration and being configured, etc.

[0031] With the development of social economy and the improvement of consumption level, people's consumption of financial products is becoming more and more common. It should be understood that financial products can include channel products, financial products, fund products, etc., and the embodiments of this application do not limit this. Users can order suitable financial products on the financial product service platform.

[0032] At present, the recommendation of financial products is based on the pre-calculation of the user's existing data by the financial product service platform. For example, the financial product service platform recommends financial products based on the user's historical browsing information, historical ordering information, credit information and other offline data. This kind of pre-calculation lacks real-time performance, cannot meet the current needs of users, and cannot quickly respond to changes in user preferences and behaviors. It is difficult to improve the accuracy and effectiveness of recommendations, which ultimately leads to a low subscription rate for financial product service platforms.

[0033] The following explanation is given by taking a credit product as an example of a financial product.

[0034] For credit products, users can order suitable credit products on the Daichao platform. It should be understood that Daichao is the abbreviation of "loan supermarket", which is a term in the Internet financial industry. Loan supermarket is similar to an online platform that brings together products from multiple lending institutions and provides borrowers with a variety of loan options.

[0035] The Daichao platform brings together various credit products from multiple lending institutions. The industry currently generally implements recommendation logic when users enter the Daichao platform to recommend credit products that can be granted or borrowed. Focusing on the borrowing process, the Daichao platform will improve the loan conversion rate through operational means such as coupons, product increase, and recommendation of new credit products.

[0036] It should be understood that the loan conversion rate is an indicator in the financial field, which is usually used to measure the ratio between the actual number of borrowers who initiate loan applications within a certain period of time and the total number of people who can initiate loan applications.

[0037] Whether it is recommending credit products, sending coupons, or increasing product credit limits, the current Daichao platform performs pre-calculations based on the platform's existing data. This type of pre-calculation lacks real-time performance and may not meet the user's current needs. The user's current behavior dynamics on the platform are only collected and not fed back to the recommendation results in real time. In summary, the current technical solutions cannot quickly respond to user preferences and behavioral changes, and it is difficult to improve the accuracy and effectiveness of recommendations, which ultimately leads to a low loan conversion rate on the Daichao platform.

[0038] In view of this, the embodiments of the present application provide a financial product recommendation method, a computer device, and a computer program product, which can determine the user's current intention based on the user's behavior information, and recommend financial products to the user in real time based on the user's intention. In this way, it can quickly respond to changes in user preferences and behavior, improve the accuracy and effect of recommendations, and thereby increase product subscription rates.

[0039] The system architecture of the embodiment of the present application is described in detail below.

[0040] See also Figure 1 , Figure 1 1 is an optional architecture diagram of a financial product recommendation system 100 provided in an embodiment of the present application, the system includes a user device 110 and a server 120, wherein the user device 110 (user device 110-1 and user device 110-2 are shown as examples) is connected to the server 120 via a network 130. It should be noted that the network 130 may be a wide area network or a local area network, or a combination of the two, and the embodiment of the present application does not limit this.

[0041] It should be noted that in the embodiment of the present application, the server may be a computer device, and the two are equivalent or interchangeable.

[0042] It should also be noted that the network 130 can also be understood as a server gateway, and the embodiments of the present application do not limit this.

[0043] In some embodiments, the user device 110 can be implemented as various types of terminals such as a smart phone, a laptop computer, a tablet computer, a desktop computer, a set-top box, a smart TV, a portable music player, a wearable device, a personal digital assistant, a dedicated messaging device, a portable gaming device, etc., but is not limited thereto.

[0044] The server 120 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The user device 110 and the server 120 may be directly or indirectly connected via wired or wireless communication, which is not limited in the embodiments of the present application.

[0045] Exemplarily, when the user device 110 is implemented as a mobile phone, the user can start a financial product service application (such as a loan application) on the mobile phone. After starting the financial product service application, the financial product can be displayed on the display interface (the display interface for displaying financial products is referred to as the product display interface in the embodiment of the present application). The mobile phone can obtain the user's behavior information on the product display interface and initiate a financial product recommendation request to the server 120. After receiving the financial product recommendation request, the server 120 can determine the user's target intent based on the user's behavior information, and determine the product that matches the target intent from multiple products pre-stored in the database 140 based on the user's target intent, and generate recommended financial product information; the server 120 can and send the recommended financial product information back to the user device 110. After the user device 110 obtains the recommended financial product information fed back by the server 120, it displays the recommended financial product information in the product display interface.

[0046] Below, the financial product recommendation method provided by the embodiment of the present application will be described in combination with the system architecture provided by the embodiment of the present application.

[0047] It should be noted that the embodiments of the present application mention financial products, including credit products, financial products, fund products, etc., and the embodiments of the present application do not impose any restrictions on this.

[0048] This application embodiment provides a financial product recommendation method, such as Figure 2 As shown, the method comprises the following steps:

[0049] S210, when the user browses the product display interface, the user device 110 obtains the user's behavior information on the product display interface;

[0050] S220, the user device 110 sends behavior information to the server 120;

[0051] S230, the server 120 determines the user's target intention based on the behavior information;

[0052] S240, the server 120 determines recommended financial product information for the user based on the target intention and the multiple financial product information;

[0053] S250, the server 120 sends recommended financial product information to the user device 110;

[0054] S260: The user device 110 displays recommended financial product information on the product display interface.

[0055] It should be understood that in step S210 , when the user device 110 detects that the user browses the financial product display interface, the user behavior information on the financial product display interface can be collected in real time.

[0056] It should be noted that when the user device 110 detects that the user starts a financial product service application and detects any touch operation (such as a click operation or a slide operation) of the user on the product display interface, it can be considered that the user is browsing the product display interface.

[0057] In some embodiments, the behavior information includes one or more of the following:

[0058] User operation information on the product display interface;

[0059] User action information;

[0060] User's expression information;

[0061] User's gaze tracking information.

[0062] The user's operation information on the product display interface may refer to the user's touch operation information on the product display interface. It should be understood that the user device may obtain the user's operation information on the product display interface through the touch display screen.

[0063] In some embodiments, the operation information may include one or more of the following: operation object, operation type, operation duration, and operation frequency. The operation object includes but is not limited to pages, cards (which can also be understood as part of the content in the page), modules, elements, links, buttons, interfaces, etc. in the product display interface. The operation type includes but is not limited to click operation, slide operation, double-click operation, two-finger or multi-finger slide operation.

[0064] Exemplarily, the operational information may be the user's clicks on the lending institution on the loan calculation page, clicks on the loan interest rate and length of time spent on it, clicks on the discount page / activity page, discount card / activity card or discount module / activity module on the marketing page and length of time the user stays on the loan agreement page, etc.

[0065] In an embodiment of the present application, the user's motion information may be the user's limb or head motion information. For example, the motion information may be the amplitude and frequency of motions such as nodding, shaking head, waving hands, frowning, etc., and the embodiment of the present application does not impose any restrictions on this.

[0066] In addition, the user's expression information may include happiness, sadness, curiosity, surprise, etc. The sight tracking information may include the position where the user's sight stays on the product display interface, the length of time the sight stays, etc.

[0067] It should be noted that the user device 110 can start the image acquisition device when the user browses the product display interface, and collect body images, facial images, or eye images of multiple users through the image acquisition device. By performing image processing on the collected multiple images, the user's action information, expression information, and eye tracking information are determined.

[0068] It should also be noted that the user's body image, facial image, or eye image involves the user's privacy information, so before the image acquisition device acquires the user's body image, facial image, or eye image, the user's authorization is required. In other words, only with the user's authorization can the user device start the image acquisition device when the user browses the product display interface, and acquire the body images, facial images, or eye images of multiple users through the image acquisition device.

[0069] In the embodiment of the present application, after the user device 110 obtains the user's behavior information, in step S220, the user device 110 may send (or report) the collected behavior information to the server 120 through the network 130. It should be understood that the server 120 may be a server of a financial product service platform, such as a server of a credit platform.

[0070] In some embodiments, the user device 110 may integrate the above-mentioned behavior information and report the integrated behavior information to the server 120 .

[0071] In other embodiments, the user device 110 may also send the above behavior information directly to the server 120, and the server 120 may integrate and process the behavior information to reduce the processing complexity of the user device.

[0072] It should be noted that the above-mentioned integration processing may include processing operations such as filtering out noise, merging repeated information, marking main behavior information, etc., and the embodiments of the present application do not limit this.

[0073] It should be understood that behavior information can be divided into primary behavior information and secondary behavior information. Among them, the importance of primary behavior information is higher than that of secondary behavior information. Exemplarily, the behavior information with high frequency of occurrence in the behavior information can be marked as primary behavior information, or the behavior with the longest appearance time in the behavior information can be marked as primary behavior, and the embodiments of the present application do not limit this. The user device 110 can mark the primary behavior information in the behavior information sent to the server 120, so that the server 120 can identify the user's intention based on the marked primary behavior information and all behavior information.

[0074] Furthermore, after receiving the user's behavior information in step S230, the server 120 may determine the user's current target intention based on the behavior information.

[0075] In the embodiment of the present application, the server 120 can input the behavior information reported by the user device 110 into the user intention recognition engine, and perform user intention recognition through the user intention recognition engine. Among them, the user intention recognition engine can rely on the classification model to classify the input behavior information through the classification model, and the output intent category is the user's current target intent.

[0076] In the embodiment of the present application, the above classification model can be a multi-classification model based on a support vector machine (SVM), or a model that implements multi-classification through a “one-to-one method” based on SVM.

[0077] For example, in order to increase the product ordering rate, the embodiment of the present application focuses on the potential intention of the user not to order the financial product. The user's intention of not ordering the financial product can also be understood as the reason why the user hesitated not to order the financial product, that is, the embodiment of the present application focuses on the user's hesitation point.

[0078] Based on this, the user's potential intentions can be divided into four categories: institution evaluation, interest rate evaluation, credit limit evaluation, and discount evaluation. It should be noted that institution evaluation can be understood as the user's low trust in the institution currently providing the product (or the trust in the institution currently providing the product does not meet the user's expectations, or the trust in the institution currently providing the product needs to be improved); interest rate evaluation can be understood as the user is dissatisfied with the interest rate of the currently displayed product (or the interest rate of the currently displayed product does not meet the user's expectations); credit limit evaluation can be understood as the user is dissatisfied with the credit limit provided by the currently displayed product (or the credit limit provided by the currently displayed product is small), and discount evaluation can be understood as the current product's discount strength does not meet the user's expectations.

[0079] Taking the above-mentioned "one-to-one method" based on SVM to achieve multi-classification as an example, the institutional assessment is marked as A, the interest rate assessment is marked as B, the credit assessment is marked as C, and the preferential assessment is marked as D. In actual applications, the server 120 can construct 6 classifiers in advance, each classifier is used to classify two intents. Specifically, these six classifiers are used to classify the intents of six situations of A:B, A:C, A:D, B:C, B:D, and C:D. In the early stage, the server 120 can use the behavior information corresponding to A:B, A:C, A:D, B:C, B:D, and C:D as a training set to train 6 classifiers. In actual applications, the behavior information sent by the user device 110 can be input into the 6 classifiers for prediction, and finally voting is performed to obtain the intent category with the highest votes, and the intent corresponding to the intent category is used as the user's target intent.

[0080] In the embodiment of the present application, after determining the user's target intent, in step S240, the server 120 can determine the recommended financial product information for the user based on the user's target intent. In other words, the server 120 can obtain recommended products that meet the user's target intent based on the identified user's target intent, thereby achieving real-time financial product recommendations.

[0081] Exemplarily, when the target intent is institutional assessment or credit limit assessment, server 120 may try to acquire (or recall) more available financial products; when the target intent is interest rate assessment, server 120 may try to acquire / recall products whose interest rates meet user needs; when the target intent is discount assessment, server 120 may try / recall more affordable marketing products.

[0082] It should be noted that the recommended financial product information may be the product information corresponding to the recommended financial product, or may be understood as the relevant information that can be displayed on the product display interface. For example, the recommended financial product information may be the introduction information (or description information) of the financial product, the institution that provides the financial product, the interest rate information of the financial product, the display method of the recommended financial product, etc., which is not limited in the embodiments of the present application.

[0083] Further, after obtaining the recommended financial product information, in step S250, the server 120 may send the recommended financial product information to the user device 110 via the network 130. Accordingly, after receiving the recommended financial product information, the user device 110 may display the recommended financial product information on the product display interface.

[0084] It can be seen that the financial product recommendation method provided in the embodiment of the present application can realize interactive financial product recommendation, wherein the user device can obtain the user's real-time behavior information for the product display interface and report it to the server. The server can better identify the user's intention based on the behavior information, thereby providing the user with real-time financial product recommendations in a more targeted manner based on the user's intention, thereby improving the product subscription rate.

[0085] In some embodiments, in step S210, the user device 110 obtains the user's behavior information for the product display interface, which may be that the user device obtains the user's behavior information for the product display interface in each first time interval according to the first time interval.

[0086] In addition, in step S220, the user equipment 110 sends the behavior information to the server 120, which may be sending the behavior information related to each first time interval to the server at the first time interval.

[0087] It should be understood that when it is detected that a user is browsing a product display interface, the user device 110 can collect behavior information of the user device 110 for the product display interface within each first time interval in units of the first time interval, and after collecting the behavior information within each time interval, send the behavior information within the time interval to the server 120.

[0088] Accordingly, the server 120 may receive a piece of behavior information at each first time interval. The server 120 may recommend financial products to the user based on the behavior information received in each first time interval, that is, the server 120 may obtain a piece of recommended financial product information in each first time interval, and may specifically send a piece of recommended financial product information to the user device 110 according to the first time interval. It should be noted that the manner in which the server 120 recommends products to the user based on the behavior information is specifically described in the above steps S230 to S250, and for the sake of brevity, it will not be repeated here.

[0089] It should be noted that the first time interval can be understood as a time window, and the first time interval can be a shorter duration, such as 2 seconds, 5 seconds or 10 seconds, and the embodiment of the present application does not impose any limitation on this.

[0090] It should also be noted that the first time interval may be a predefined or preconfigured parameter, or may be a parameter dynamically configured by the server 120 for the user equipment 110, and the implementation of this application does not impose any limitation on this.

[0091] Through the financial product recommendation method provided in the embodiment of the present application, the user device 110 collects behavior information within each time interval according to the first time interval, thereby realizing real-time collection of behavior information. Correspondingly, the server 120 can identify the user's intention based on the real-time behavior information and make real-time recommendations.

[0092] In one embodiment of the present application, refer to Figure 3 As shown, in step S260, the user device 110 displays the recommended financial product information on the product display interface, which can also be achieved in the following ways:

[0093] S2601, the user equipment 110 obtains animation rendering parameters;

[0094] S2602: The user device 110 displays the recommended financial product information and the animation effects corresponding to the recommended financial product information on the product display interface based on the animation rendering parameters.

[0095] The animation effects may be shaking, enlarging, flipping, flashing, etc. The animation rendering parameters may include but are not limited to: shaking times, shaking duration, enlarging ratio, enlarging duration, flipping direction, flashing color, etc.

[0096] In some embodiments, the animation rendering parameters may be predefined parameters. For example, the animation rendering parameters may be implemented by pre-saving a response code, table, or other methods that can be used to indicate the animation rendering parameters in the user device 110 .

[0097] In other embodiments, the animation rendering parameters may also be parameters dynamically configured by the server 120 for the user device 110. For example, when the server 120 sends recommended financial product information to the user device 110, it may also send the animation rendering parameters corresponding to the recommended financial product information to the user device 110.

[0098] Through the financial product recommendation method provided in the embodiment of the present application, when the user device 110 displays the recommended financial product information on the product display interface, it can also be displayed in combination with animation special effects, thereby guiding the user to click on the financial product corresponding to the recommended financial product information and improving the financial product subscription rate.

[0099] In one embodiment of the present application, in S230, the server 120 determines the user's target intention based on the behavior information in multiple ways, which are introduced below using way #A and way #B as examples.

[0100] Method #A: Reference Figure 4A As shown, in step S230, the server 120 determines the user's target intention based on the behavior information, which can be achieved in the following ways:

[0101] S2301A: The server 120 uses the behavior information as input of the first model, and outputs a first intent category through the first model, wherein the first intent category is used to indicate the target intent of the user.

[0102] It should be noted that the first model is used to classify the input behavior information and output the intent category to which the behavior information belongs. The first model can be a multi-classification model based on SVM, or a multi-classification model implemented by the "one-to-one method" based on SVM, which is not limited in the embodiments of the present application.

[0103] It should be understood that the server 120 can input all the received behavior information into the first model, make predictions through the first model, and determine the user's target intention.

[0104] In the embodiment of the present application, the user's potential intentions are divided into four categories: institution evaluation, interest rate evaluation, credit evaluation, and discount evaluation. The first intention category output by the first model can be one or more of the above four categories of potential intentions, and the embodiment of the present application does not limit this.

[0105] Exemplarily, when the first model is a multi-classification model based on the "one-to-one method" of SVM, the server 120 can input the behavior information sent by the user device 110 into 6 classifiers for prediction, and finally vote to obtain one or more intent categories with the highest votes to obtain the first intent category.

[0106] It should be noted that the potential intention can be understood by referring to the relevant description in the above embodiment, and for the sake of brevity, it will not be repeated here.

[0107] Mode #B: The behavior information reported by the user device 110 includes primary behavior information and secondary behavior information; the importance of the primary behavior information is greater than the importance of the secondary behavior information. It should be noted that the primary behavior information and the secondary behavior information can refer to the description in the above embodiment, and for the sake of brevity, they will not be repeated here.

[0108] refer to Figure 4B As shown, in step S230, the server 120 determines the user's target intention based on the behavior information, which can be achieved in the following ways:

[0109] S2301B, using the primary behavior information as input to the second model, and using the primary behavior information and the secondary behavior information as input to the third model;

[0110] S2302B. Within the first time period, if the second intent category output by the second model and the third intent category output by the third model are received, the target intent is the intent indicated by the third intent category; if the second intent category output by the second model is received and the third intent category output by the third model is not received, the target intent is the intent indicated by the second intent category.

[0111] It should be noted that the second model and / or the third model is used to classify the input behavior information and output the intent category to which the behavior information belongs.

[0112] It can be understood that the first model, the second model, and the third model are all the same model, or in other words, the first model, the second model, and the third model are the same model. The first model to the third model are all used to determine the intent category to which the behavior information belongs based on the input behavior information.

[0113] In the embodiment of the present application, the user's potential intentions are divided into four categories: institution evaluation, interest rate evaluation, credit limit evaluation, and discount evaluation. The second intention category output by the second model and the third intention category output by the third model can be one or more of the above four categories of potential intentions, and the embodiment of the present application does not limit this.

[0114] In the embodiment of the present application, in order to ensure that the response time is controllable, the server 120 can use the main behavior information in the behavior information as the input of the second model alone, and use all the behavior information (i.e., including the main behavior information and the secondary behavior information) as the input of the third model. In other words, the server 120 can input the main behavior information and all the behavior information into the model in parallel for prediction, that is, the second model and the third model perform predictions in parallel and independently. It can be understood that the input parameters of the second model only include the main behavior information, so the prediction time of the second model will be shorter.

[0115] It should be understood that if both the second model and the third model output the intent category within the first time period, the intent category input by the third model predicted based on all behavior information is used as the target intent of the end user. If only the second model outputs the intent category within the first time period, the intent category output by the second model is used as the target intent of the end user. In this way, the response time of the model can be improved, and financial product recommendations can be provided to users more quickly.

[0116] In an embodiment of the present application, in step S220, the user device 110 sends the behavior information to the server 120, and also sends the user attribute information to the server 120. The user attribute information may include one or more of the following:

[0117] User location information;

[0118] User identity information;

[0119] User device information;

[0120] User credit information.

[0121] It should be noted that the location information of the user may be latitude and longitude information, or identification information of the cell where the user resides, which is not limited in the embodiments of the present application. User identity information may be information that uniquely identifies the user, such as a user ID, a user name, and a user's mobile phone number. Device information may be identification information of the user's device, a model of the user's device, and other information. User credit information may be a user credit score or risk score information obtained based on the user's historical ordering information.

[0122] It should be noted that the user attribute information and the behavior information together determine the user's target intention. That is, in step S230, the server 120 determines the user's target intention based on the behavior information, which can also be implemented in the following way: the server 120 determines the user's target intention based on the behavior information and the user attribute information. Specifically, the server 120 can use the behavior information and the user attribute information as the input of the model, and determine the user's target intention through the model, where the model can be one or more of the above-mentioned first model, second model, and third model.

[0123] In some embodiments, the server 120 may input the behavior information and user attribute information reported by the user device 110 into the first model together, and output the user's target intention through the first model.

[0124] In other embodiments, the server 120 may input the main behavior information and user attribute information in the behavior information reported by the user device 110 into the second model together, and simultaneously input all the behavior information and user attribute information reported by the user device 110 into the third model together. If both the second model and the third model output the intent category within the first duration, the intent category input by the third model predicted based on all the behavior information is used as the target intent of the end user. If only the second model outputs the intent category within the first duration, the intent category output by the second model is used as the target intent of the end user.

[0125] It should be understood that in this embodiment, the training data sets of the first model to the third model may include behavior information and user attribute information. That is, in this implementation, the first model to the third model can combine the user's behavior information and user attribute information to predict the user's intention, which can improve the accuracy of the prediction and achieve accurate prediction of the intention.

[0126] The following describes in detail how the server 120 determines the recommended financial product information based on the target intent and the multiple financial product information.

[0127] In one embodiment of the present application, refer to Figure 5 As shown, in step S240, the server 120 determines the recommended financial product information for the user based on the target intention and the multiple financial product information, which can be achieved in the following ways:

[0128] S2401, the server 120 determines a plurality of first recommended products matching the target intent from a plurality of financial products indicated by a plurality of financial product information;

[0129] S2402, the server 120 determines a second recommended product from the plurality of first recommended products based on the user's credit information and / or relevant information of each of the plurality of first recommended products;

[0130] S2403. The server 120 determines recommended financial product information based on the second recommended product.

[0131] It should be understood that the multiple financial product information may include the product information of all financial products included in the list of financial products maintained by the server 120 or the financial product resource pool. The server 120 may obtain / recall the first recommended product that matches or meets the target intent from the financial product resource pool based on the user's target intent. For example, when the target intent is institutional assessment or quota assessment, the server 120 may try to obtain / recall more available first recommended products from the product resource pool; when the target intent is interest rate assessment, the server 120 may try to obtain / recall the first recommended product whose interest rate meets the user's interest rate requirement from the product resource pool; when the target intent is discount assessment, the server 120 may try to obtain / recall a more affordable first recommended product from the product resource pool.

[0132] It should be noted that after acquiring / recalling the first recommended products according to the target intent, the server 120 may filter the first recommended products, remove the recommended products that do not meet the requirements, and obtain the second recommended products that meet the requirements.

[0133] In an embodiment of the present application, the server 120 may filter the acquired / recalled first recommended products in combination with the current user's credit information and / or relevant information of each first recommended product to obtain the second recommended product.

[0134] It should be understood that the user credit information may be a user credit score obtained based on the user's historical order information, or risk score information.

[0135] For example, the server 120 may remove the first recommended products whose user credit information does not meet the credit requirements of the first recommended products, and obtain the second recommended products.

[0136] It should be noted that the relevant information of each first recommended product may include one or more of the following:

[0137] Maintenance information of the first recommended product (e.g., whether it is in a credit maintenance state or a loan maintenance state);

[0138] Number of recommendations for the first recommended product;

[0139] The recommendation times threshold of the first recommended product;

[0140] Information on the number of people granted credit for the first recommended product;

[0141] The credit limit of the first recommended product;

[0142] The credit limit information of the first recommended product;

[0143] The credit limit value of the first recommended product;

[0144] The expiration date information of the first recommended product;

[0145] The first recommended product is applicable to the device information.

[0146] For example, the server 120 may eliminate products from the first recommended products according to the following rules to obtain the second recommended products:

[0147] Eliminate the products that have been discontinued from the first recommended products;

[0148] Eliminate products that are under credit maintenance or loan maintenance from the first recommended products;

[0149] Eliminate the products that have reached the recommended times threshold value on that day among the first recommended products;

[0150] Eliminate the products that have reached the credit number threshold for the day from the first recommended products;

[0151] Eliminate the products that have reached the daily quota threshold among the first recommended products;

[0152] Eliminate the first recommended products that have been approved, recommended and not expired

[0153] Eliminate products from the first recommended products that cannot be displayed on the current user's device.

[0154] It should be noted that after obtaining the second recommended product, the server 120 may determine the recommended financial product information to be sent to the user device 110 based on the second recommended product.

[0155] In a possible implementation, the server 120 may directly obtain the product information of the second recommended product, and send the product information of the second recommended product as the recommended financial product information to the user device 110. Accordingly, the user device 110 may display the second recommended product in the product display interface based on the recommended financial information.

[0156] In another possible implementation, the server 120 may select a second recommended product whose priority satisfies a condition (e.g., the highest priority, or a priority higher than a preset priority) from the second recommended products to obtain a target recommended product. Then, the product information of the target recommended product is sent to the user device 110 as recommended financial product information. Accordingly, the user device may display the target recommended product in the product display interface based on the received recommended financial product information.

[0157] It should be noted that the priority of the recommended product is determined based on the life cycle value (LTV) information of the recommended product and / or the targeted recommendation information. Among them, the LTV information of the recommended product can be determined based on the sum of all economic benefits of the product. The targeted recommendation information can be that the operation sets a targeted recommendation for the recommended product.

[0158] In actual applications, LTV information can be expressed in the form of LTV scores. Products with higher LTV scores have higher priorities, and vice versa. Products with lower LTV scores have lower priorities. In addition, products with targeted recommendation information have higher priorities than products without targeted recommendation information. For example, the server 120 acquires / recalls products including 5 financial products A, B, C, D, and E, of which B and C are products recalled under targeted operational intervention. In the recall configuration, the priority of product B configuration is higher than that of product C. For the remaining products A, D, and E, they are ranked from high to low according to their LTV scores as A, E, and D. The final ranking is B, A, E, C, and D.

[0159] In one example, if the number of second recommended products determined in step S2402 includes multiple products, the server 120 may sort the multiple second recommended products in descending order of priority of the second recommended products. Accordingly, in step S2403, the server 120 determines the recommended financial product information based on the second recommended products, which may be that the server 120 selects the first-ranked second recommended product from the sorted multiple second recommended products, and sends the product information of the first-ranked second recommended product to the user device 110.

[0160] In another possible implementation, after obtaining the second recommended product, the server 120 may further determine the recommended financial product information to be ultimately sent to the user device 110 in combination with the targeted delivery information of the operation configuration.

[0161] Specifically, refer to Figure 6 As shown, in step S2403, the server 120 determines the recommended financial product information based on the second recommended product, which can also be implemented in the following ways:

[0162] S24031. The server 120 determines whether the user meets the conditions for targeted delivery;

[0163] S24032: If the user meets the targeted delivery condition, the server 120 adds the targeted delivery product associated with the targeted delivery condition to the second recommended product to obtain an updated second recommended product;

[0164] S24033. The server 120 determines recommended financial product information based on the updated second recommended product.

[0165] It should be noted that targeted advertising is an advertising strategy that accurately pushes recommended content to target user groups based on user-specific attributes, interests and behaviors. Targeted advertising generally configures targeted financial products according to population packages.

[0166] In the embodiment of the present application, the server 120 can determine whether the user hits the crowd package, that is, whether the user meets the targeted delivery condition, based on the user attribute information reported by the user device 110. If the user meets the targeted delivery condition, the targeted delivery product associated with the targeted delivery condition is added to the second recommended product to obtain an updated second recommended product.

[0167] Further, the server 120 may determine the recommended financial product information based on the updated second recommended product. In one implementation, the server 120 may directly obtain the product information of the updated second recommended product, and send the updated product information of the second recommended product to the user device 110 as the recommended financial product information. In another implementation, the server 120 may select a second recommended product whose priority satisfies the conditions (e.g., the highest priority, or the priority is higher than the preset priority) from the updated second recommended products to obtain the target recommended product. The product information of the target recommended product is then sent to the user device 110 as the recommended financial product information. It should be noted that the priority of the recommended product is determined based on the LTV information and / or the directional recommendation information of the recommended product. For the specific content, please refer to the description above, and for the sake of brevity, it will not be repeated here.

[0168] It should be noted that, before recommending financial products, the server 120 needs to make an admission judgment on the acquired product to be recommended (eg, the second recommended product mentioned above, the updated second recommended product).

[0169] The embodiments of this application are Figure 5 The method shown in the figure is taken as an example, that is, the product to be recommended is the second recommended product as an example to illustrate the process of admission judgment. Figure 5 In the method shown, in step S2403, the server 120 determines the recommended financial product information based on the second recommended product, which can be implemented in the following manner:

[0170] The server 120 sends the user identity information to the provider of the second recommended product; the user identity information is used by the provider to determine whether to allow the user to use the financial product provided by the provider;

[0171] The server 120 receives access information sent by the provider of the second recommended product; the access information is used to indicate whether the use of the financial product provided by the provider is allowed;

[0172] The server 120 obtains the second recommended product that the corresponding access information indicates is allowed to be used from the second recommended product, and obtains the third recommended product;

[0173] The server 120 determines the recommended financial product information based on the third recommended product.

[0174] It should be noted that the user identity information may be information that uniquely identifies the user identity, such as a user identity ID, a user name, and a user mobile phone number. The user identity information may also be information obtained by performing MD5 processing on one or more of the user identity ID, user name, and user mobile phone number. This embodiment of the application does not limit this.

[0175] In addition, the provider of the second recommended product can also be understood as the channel party behind the second recommended product.

[0176] In the embodiment of the present application, the server 120 may send the user identity information to the provider of the second recommended product so that the provider can make an access judgment on the current user. The access judgment refers to whether the user is allowed to use the second recommended product. In other words, the provider can judge whether the user is allowed to use the second recommended product based on the user identity information sent by the server 120.

[0177] Furthermore, the provider of the second recommended product may send access information to the server 120, and the access information may carry the access judgment result. That is, the access information may indicate whether the access is passed or not. If the access is passed, the current user is allowed to use the second recommended product, and if the access is not passed, the current user is not allowed to use the second recommended product.

[0178] It should be noted that when the number of second recommended products includes multiple ones, the server 120 can send the user's identity information to the providers of each second recommended product, so that each product provider can make an access judgment on the current user. Accordingly, the server 120 can receive the access information corresponding to each second recommended product. In this way, the server 120 can obtain the second recommended product whose access information indicates that it has passed, and obtain the third recommended product.

[0179] In a possible implementation, the server 120 may directly obtain product information of the third recommended product, and send the product information of the third recommended product as recommended financial product information to the user device 110. Accordingly, the user device 110 may display the third recommended product in the product display interface based on the recommended financial product information.

[0180] In another possible implementation, the server 120 may select a third recommended product whose priority satisfies the condition (e.g., the highest priority, or a priority higher than a preset priority) from the third recommended products to obtain a target recommended product. The product information of the target recommended product is then sent to the user device 110 as recommended financial product information. Accordingly, the user device may display the target recommended product in the product display interface based on the received recommended financial product information. It should be noted that the priority of the recommended product is determined based on the LTV information and / or directional recommendation information of the recommended product. For the specific content, please refer to the description above, and for the sake of brevity, it will not be repeated here.

[0181] In summary, the embodiment of the present application provides an interactive financial product recommendation method, in which the user device can obtain and report the user's behavior information on the product display interface, and the server identifies the user's target intention through the behavior information, and makes real-time recommendations based on the user's target intention. In this way, real-time recommended products can be provided to users in a more targeted manner, ultimately increasing the product subscription rate.

[0182] The following is a detailed description of the method provided in the embodiment of the present application using a credit product recommendation scenario.

[0183] First, key terms in the credit product recommendation scenario are introduced.

[0184] Interactive recommendation: Interactive recommendation is an interactive real-time recommendation product module that mainly makes recommendations in an interactive way by understanding user needs.

[0185] Daichao: Daichao is the abbreviation of "loan supermarket", which is a term in the Internet finance industry. A loan supermarket is similar to an online platform that brings together products from multiple lending institutions and provides borrowers with a variety of loan options.

[0186] Loan conversion rate: Loan conversion rate is an indicator in the financial field, which is usually used to measure the ratio between the actual number of borrowers who initiate loan applications within a certain period of time and the total number of people who can initiate loan applications.

[0187] The present application embodiment provides a credit product recommendation method, specifically, referring to Figure 7As shown in the flow diagram, when a user enters the Daichao platform, credit products are distributed to the user according to the pre-calculation and the relevant operation positions are displayed. When the user generates a trigger operation (which can also be understood as a touch operation, such as sliding, clicking, staying, etc.) on the Daichao page, the user's behavior information is recorded in the feature information, which includes page code, card code, key elements, user actions, user features, product information, etc. The feature information is then reported to the Daichao server (i.e., the server mentioned above), and the Daichao server identifies the user's intention based on the feature information and executes the real-time recommendation logic. The Daichao front end (i.e., the user device mentioned above) receives the recommended data returned by the server, which may be a credit limit increase card, coupon card, credit card, loan card, etc. When the user returns to the Daichao homepage, the front end renders the card content in real time and displays it to the user, and guides the user to click through dynamic effects such as card shaking, thereby improving the loan conversion rate.

[0188] Trigger action description: This solution focuses on the user's sliding, clicking, and dwelling time (2s / 5s / 10s) on the core pages and front-end components of Daichao, such as the user's clicks on the lending institution and the clicks and dwell time on the loan interest rate on the loan calculation page, the user's clicks on the coupon module and activity module on the marketing page and the dwell time of the user, and the dwell time of the user on the loan agreement page, etc.

[0189] Feature information processing and reporting: The Daichao front end will integrate and process the user-triggered actions described above. Figure 8 As shown in the reporting process diagram, the Daichao front-end can use 5 seconds as a time window to find the user's most frequently triggered action and mark it as the main triggering action in the current time window. After the integration process is completed, the front-end will report the integrated user behavior, current latitude and longitude, device information, etc. as feature information to the Daichao server through the HTTP interface.

[0190] Recommendation process design: After the Daichao server obtains the reported feature information, it first identifies the user's intention, recalls the recommended products according to different identification results, and then filters the recalled products according to a series of conditions. The filtered products are sorted according to specific rules. Some products also need to request the channel interface for access judgment. Finally, the products that have passed the access and ranked first are returned to the front end.

[0191] The following is an explanation of each link.

[0192] User intention recognition: The Daichao service inputs the reported feature information into the user intention recognition engine. The engine relies on SVM to achieve multiple classifications through multiple "one-to-one methods". The final winning classification is the user's current behavioral intention. At present, we divide the user's behavioral intention into loan institution evaluation, loan interest rate evaluation, loan amount evaluation, and loan discount evaluation.

[0193] The following is an implementation of SVM multi-classification. According to the user's potential behavioral intention, the loan institution evaluation is marked as A, the loan interest rate evaluation is marked as B, the loan amount evaluation is marked as C, and the loan discount evaluation is marked as D. In the early stage, the feature data corresponding to A:B, A:C, A:D, B:C, B:D, and C:D are used as training sets, and then 6 training models are obtained. The behavioral features in the subsequent feature information are predicted on the 6 models respectively, and finally voted, and the one with the highest votes is obtained as the acceptance result.

[0194] refer to Fig. 9 In the user intent recognition process diagram shown in the figure, in order to ensure that the response time is controllable, the main action will be taken out separately, and the main action and all trigger actions will be predicted by the SVM model in parallel. If only the main action is used as an input parameter, the corresponding model prediction time will be shorter. If both parties return the acceptance result within the response time of 500ms, the acceptance result returned by all trigger actions shall prevail, otherwise the acceptance result returned by the main action shall prevail.

[0195] The following describes the product recall process.

[0196] refer to Fig.10 The product recall process diagram is shown in the figure. The Daichao service end can recall products in different dimensions according to different behavioral intentions of users. For example, the two behavioral intentions of borrowing institution assessment and borrowing amount assessment will try to recall more available credit products; borrowing interest rate assessment will try to recall credit products with lower interest rates; and borrowing discount assessment will try to recall more affordable marketing products. After the Daichao service end recalls credit products according to behavioral intention classification, it will combine the current user's internal risk score and credit score as well as the targeted delivery of the operation configuration (generally, targeted credit products are configured according to population packages) to form the final collection of recalled products.

[0197] The product filtering process is described below.

[0198] It should be understood that after the Daichao service end obtains the recalled product set, it can judge the recalled products according to a series of rules and remove the products that do not meet the conditions from the recalled product set. Among them, the filtering rules are as follows:

[0199] A. Determine whether there are any discontinued products in the recalled product set and remove the discontinued products;

[0200] B. Determine whether there are products in the recalled product set that are under credit maintenance or loan maintenance, and remove the products in such conditions;

[0201] C. Determine whether there are products in the recalled product set that have reached the upper limit of the number of recommendations for the day, and remove the products that have reached the upper limit of the number of recommendations for the day;

[0202] D. Determine whether there are products in the recalled product set that have reached the upper limit of the number of people granted credit on that day, and remove the products that have reached the upper limit of the number of people granted credit on that day;

[0203] E. Determine whether there are products in the recalled product set that have reached the upper limit of the daily loan amount, and remove the products that have reached the upper limit of the daily loan amount;

[0204] F. Determine whether there are any products in the recalled product collection that have been approved, recommended, and not expired, and remove products that have been approved, expired, or recommended, expired;

[0205] G. Determine whether there are products in the recalled product set that cannot be displayed on the current device, and remove the products that cannot be displayed on the current device.

[0206] The sorting process is described below.

[0207] refer to Fig.11 The flow chart of the sorting method shown in the figure shows that after obtaining the recalled product set that has passed the filter, the products in the set are sorted according to the rule of first sorting by operation orientation and then sorting by LTV. For example: There are 5 credit products A, B, C, D, and E in the recalled products, among which B and C are products recalled under the intervention of operation orientation. In the recall configuration, product B is configured in the first place and product C is configured in the fourth place; for the remaining products A, D, and E, they are sorted from high to low according to the LTV scores of the products as A, E, and D. The final ranking is B, A, E, C, and D.

[0208] The admission process is described below.

[0209] refer to Fig.12 The access method flow chart is shown in the figure. A round of pre-screening is performed based on the user's latitude and longitude, device information, ID number, etc. in the feature information. The specific steps of the pre-screening include: user age baffle, mobile phone number segment baffle, region baffle, and offline population package baffle. After passing the pre-screening baffle, the user's ID card information and mobile phone number information are processed by MD5 and transmitted to the channel party behind the corresponding credit product. After the channel party passes the access, the product currently ranked first is selected and returned to the front end.

[0210] In summary, the embodiment of the present application provides a solution for improving the conversion rate of loan-to-borrowing based on interactive recommendation. The solution includes: recording characteristic information and reporting it to the loan-to-borrowing service end according to the user's triggering actions such as sliding, clicking, and staying on each page of the loan-to-borrowing platform. The loan-to-borrowing service end identifies the user's intention by analyzing the characteristic information, and then performs real-time recommendations: such as recommending new credit products to users who have no borrowing products, providing a credit limit increase channel for users with insufficient credit limits, recommending low-interest credit products to users with excessively high interest rates, and providing platform endorsements for users with low trust in institutions, etc. Respond to the user's behavioral dynamics on the loan-to-borrowing page more quickly, provide users with more accurate recommendation results, and then guide users to click in combination with card animations, ultimately achieving the purpose of improving the loan conversion rate of the loan-to-borrowing platform.

[0211] An embodiment of the present application also provides a financial product recommendation device, referring to Fig.13 As shown, the financial product recommendation device includes:

[0212] The receiving unit 1310 is configured to receive behavior information sent by a user device; the behavior information is the behavior information of the user when browsing the product display interface;

[0213] The processing unit 1320 is configured to determine the target intention of the user based on the behavior information;

[0214] The processing unit 1320 is further configured to determine recommended financial product information for the user based on the target intention and the plurality of financial product information;

[0215] The sending unit 1330 is configured to send the recommended financial product information to the user equipment, where the recommended financial product information is used to be displayed on the product display interface.

[0216] In some embodiments, the processing unit 1320 is configured to use the behavior information as an input of a first model and output a first intent category through the first model; the first intent category is used to indicate the user's target intention, and the first model is used to classify the input behavior information and output the intent category to which the behavior information belongs.

[0217] In some embodiments, the behavior information includes primary behavior information and secondary behavior information; the importance of the primary behavior information is greater than the importance of the secondary behavior information;

[0218] The processing unit 1320 is configured to use the main behavior information as the input of the second model, and to use the main behavior information and the secondary behavior information as the input of the third model; the second model and / or the third model is used to classify the input behavior information and output the intention category to which the behavior information belongs; within a first time period, if the second intention category output by the second model and the third intention category output by the third model are received, then the target intention is the intention indicated by the third intention category; if the second intention category output by the second model is received and the third intention category output by the third model is not received, then the target intention is the intention indicated by the second intention category.

[0219] In some embodiments, the processing unit 1320 is further configured to determine a plurality of first recommended products that match the target intent from a plurality of financial products indicated by a plurality of financial product information; determine a second recommended product from the plurality of first recommended products based on user credit information and / or relevant information of each first recommended product in the plurality of first recommended products; and determine the recommended financial product information based on the second recommended product.

[0220] In some embodiments, the relevant information of each first recommended product includes one or more of the following:

[0221] maintenance information of the first recommended product;

[0222] The number of times the first recommended product is recommended;

[0223] A recommendation times threshold value of the first recommended product;

[0224] Information on the number of people who have been granted credit for the first recommended product;

[0225] The credit limit value of the first recommended product;

[0226] The credit limit information of the first recommended product;

[0227] the credit limit value of the first recommended product;

[0228] The validity period information of the first recommended product;

[0229] Device information to which the first recommended product is applicable.

[0230] In some embodiments, the processing unit 1320 is further configured to determine whether the user meets the targeted delivery conditions; if the user meets the targeted delivery conditions, the targeted delivery product associated with the targeted delivery conditions is added to the second recommended product to obtain an updated second recommended product; based on the updated second recommended product, the recommended financial product information is determined.

[0231] In some embodiments, the sending unit 1330 is further configured to send user identity information to the provider of the second recommended product; the user identity information is used by the provider to determine whether to allow the user to use the financial product provided by the provider;

[0232] The receiving unit 1310 is further configured to receive access information sent by the provider of the second recommended product; the access information is used to indicate whether the use of the financial product provided by the provider is allowed;

[0233] The processing unit 1320 is further configured to obtain the second recommended product that is allowed to be used as indicated by the corresponding access information from the second recommended product to obtain a third recommended product; and determine the recommended financial product information based on the third recommended product.

[0234] In some embodiments, the recommended financial product information is product information of a target recommended product, and the target recommended product is any one of the following:

[0235] The second recommended product that meets the priority conditions among the second recommended products;

[0236] The second recommended product that meets the priority conditions among the updated second recommended products;

[0237] The third recommended products that meet the priority conditions among the third recommended products;

[0238] The priority is determined based on the full life cycle value information and / or targeted recommendation information of the recommended product.

[0239] In some embodiments, the behavior information includes one or more of the following:

[0240] Operation information of the user on the product display interface;

[0241] User action information;

[0242] User's expression information;

[0243] User's gaze tracking information.

[0244] In some embodiments, the operation information includes one or more of the following:

[0245] Operation object, operation type, operation duration, and operation frequency.

[0246] In some embodiments, the receiving unit 1310 is further configured to receive user attribute information sent by the user device; wherein the input parameters of the model also include the user attribute information, and the model includes one or more of a first model, a second model, and a third model;

[0247] The user attribute information includes one or more of the following:

[0248] User location information;

[0249] User identity information;

[0250] User device information;

[0251] User credit information.

[0252] Those skilled in the art should understand that the relevant description of the above-mentioned financial product recommendation device in the embodiment of the present application can be understood by referring to the relevant description of the financial product recommendation method in the embodiment of the present application.

[0253] Fig.14 It is a schematic structural diagram of a computer device provided in an embodiment of the present application. It should be understood that the computer device 1400 can be used to implement the corresponding functions implemented by the server in the above method. For the sake of brevity, it will not be repeated here. Fig.14 The computer device 1400 shown includes a processor 1410, which can call and run a computer program from a memory to implement the method in the embodiment of the present application.

[0254] Alternatively, if Fig.14 As shown, the server 1400 may further include a memory 1420. The processor 1410 may call and run a computer program from the memory 1420 to implement the method in the embodiment of the present application.

[0255] The memory 1420 may be a separate device independent of the processor 1410 , or may be integrated into the processor 1410 .

[0256] Alternatively, if Fig.14 As shown, the server 1400 may further include a transceiver 1430, and the processor 1410 may control the transceiver 1430 to communicate with other devices, specifically, may send information or data to other devices, or receive information or data sent by other devices.

[0257] The transceiver 1430 may include a transmitter and a receiver. The transceiver 1430 may further include an antenna, and the number of the antennas may be one or more.

[0258] The present application also provides a computer storage medium, specifically a computer-readable storage medium, on which computer instructions are stored. When the computer storage medium is located in a computer device, the computer instructions are executed by a processor to implement any step of the above-mentioned financial product recommendation method in the present application.

[0259] An embodiment of the present application also provides a computer program product, including computer program instructions.

[0260] The computer program product can be applied to the computer device in the embodiments of the present application, and the computer program instructions enable the computer to execute the corresponding processes implemented by the server in the various methods of the embodiments of the present application. For the sake of brevity, they will not be repeated here.

[0261] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0262] The units described above as separate components may or may not be physically separated, and the components displayed 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 scheme of this embodiment.

[0263] In addition, all functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may be separately configured as a unit, or at least two units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0264] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium, which, when executed, executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, disks or optical disks.

[0265] Alternatively, if the above-mentioned integrated unit of the present application is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiment of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0266] It should be noted that the technical solutions described in the embodiments of the present application can be combined arbitrarily without conflict.

[0267] 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 financial product recommendation method, characterized in that: The method comprises: Receiving behavior information sent by a user device, wherein the behavior information is behavior information of a user when browsing a product display interface; Determining the target intention of the user according to the behavior information; Determining recommended financial product information for the user based on the target intention and the plurality of financial product information; The recommended financial product information is sent to the user equipment, and the recommended financial product information is used to be displayed on the product display interface.

2. The method according to claim 1, characterized in that Determining the user's target intention according to the behavior information includes: The behavior information is used as the input of the first model, and the first intention category is output through the first model; the first intention category is used to indicate the target intention of the user, and the first model is used to classify the input behavior information and output the intention category to which the behavior information belongs.

3. The method according to claim 1, characterized in that The behavior information includes primary behavior information and secondary behavior information; the importance of the primary behavior information is greater than the importance of the secondary behavior information; Determining the user's target intention according to the behavior information includes: The main behavior information is used as input of the second model, and the main behavior information and the secondary behavior information are used as input of the third model; the second model and / or the third model is used to classify the input behavior information and output the intention category to which the behavior information belongs; Within the first time period, if the second intent category output by the second model and the third intent category output by the third model are received, the target intent is the intent indicated by the third intent category; within the first time period, if the second intent category output by the second model is received and the third intent category output by the third model is not received, the target intent is the intent indicated by the second intent category.

4. The method according to claim 1, characterized in that: The determining, based on the target intention and the plurality of financial product information, recommended financial product information for the user includes: Determining, from among the plurality of financial products indicated by the plurality of financial product information, a plurality of first recommended products matching the target intent; Determine a second recommended product from the plurality of first recommended products according to the user credit information and / or relevant information of each first recommended product in the plurality of first recommended products; The recommended financial product information is determined according to the second recommended product.

5. The method according to claim 4, characterized in that The relevant information of each first recommended product includes one or more of the following: maintenance information of the first recommended product; The number of times the first recommended product is recommended; A recommendation times threshold value of the first recommended product; Information on the number of people who have been granted credit for the first recommended product; The credit limit value of the first recommended product; The credit limit information of the first recommended product; the credit limit value of the first recommended product; The validity period information of the first recommended product; Device information to which the first recommended product is applicable.

6. The method according to claim 4, characterized in that The determining the recommended financial product information according to the second recommended product includes: Determining whether the user meets the targeted delivery conditions; When the user meets the targeted delivery condition, adding the targeted delivery product associated with the targeted delivery condition to the second recommended product to obtain an updated second recommended product; The recommended financial product information is determined according to the updated second recommended product.

7. The method according to claim 4, characterized in that The determining the recommended financial product information according to the second recommended product includes: Sending the user identity information to the provider of the second recommended product; the user identity information is used by the provider to determine whether to allow the user to use the financial product provided by the provider; receiving access information sent by the provider of the second recommended product; the access information is used to indicate whether the use of the financial product provided by the provider is allowed; Acquire, from the second recommended product, the second recommended product that the corresponding access information indicates is allowed to be used, to obtain a third recommended product; Based on the third recommended product, the recommended financial product information is determined.

8. The method according to claim 4, 6 or 7, characterized in that: The recommended financial product information is product information of a target recommended product, and the target recommended product is any one of the following: The second recommended product that meets the priority conditions among the second recommended products; The second recommended product that meets the priority conditions among the updated second recommended products; The third recommended products that meet the priority conditions among the third recommended products; The priority is determined based on the full life cycle value information and / or targeted recommendation information of the recommended product.

9. The method according to any one of claims 1 to 7, characterized in that: The behavior information includes one or more of the following: Operation information of the user on the product display interface; User action information; User's expression information; User's gaze tracking information.

10. The method according to claim 9, characterized in that The operation information includes one or more of the following: Operation object, operation type, operation duration, and operation frequency.

11. The method according to claim 2 or 3, characterized in that: The method further comprises: Receiving user attribute information sent by the user equipment; wherein the input parameters of the model also include the user attribute information, and the model includes one or more of a first model, a second model, and a third model; The user attribute information includes one or more of the following: User location information; User identity information; User device information; User credit information.

12. A computer device, characterized in that: The method comprises a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method according to any one of claims 1 to 11.

13. A computer program product, characterized in that The method comprises computer program instructions, which, when executed by a processor, implement the method according to any one of claims 1 to 11.

Images