Financial product recommendation method and device, electronic equipment and program product

Abstract

The invention discloses a financial product recommendation method and device, electronic equipment and a program product, and relates to the field of financial science and technology or other related fields, and the method comprises the steps: obtaining user features of a target user and financial product features, dividing the preprocessed features into N parts, the features are processed multiple times based on a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, and the Mth processing comprises the steps of inputting the first to Mth features into the algorithm to obtain a recommendation list, and comparing financial products browsed or purchased by a target user in the (M + 1) th feature with the list to obtain a correct recommendation number, and determining a target recommendation algorithm from the collaborative filtering algorithm based on the plurality of correct recommendation quantities, inputting the N features into the target recommendation algorithm to obtain a target recommendation list, and sending recommendation information to the target user. The technical problem of low yield of financial products caused by low promotion efficiency of a financial institution on configured financial products in related technologies is solved.

Description

Technical Field

[0001] The present invention relates to the field of financial technology or other related fields, and in particular, to a method and device for recommending financial products, an electronic device, and a program product. Background Art

[0002] With the development of financial technology, financial institutions continue to expand online businesses such as wealth management, loans, and payment. Through data analysis, financial institutions continue to explore users' online behaviors, tap into users' potential needs, and meet the financial needs of different types of users through personalized service recommendations. Among them, payroll services are an important way for financial institutions to obtain stable customer sources in batches. Financial institutions provide payroll services for enterprises and units, recommend financial products to users when wages arrive, and thus increase the yield of financial institutions.

[0003] In the related art, financial institutions usually recommend financial products to users in the form of text messages when their salaries are credited. However, existing recommendation strategies have many limitations. For example, existing recommendation strategies cannot make accurate recommendations to users, and it is difficult to attract users from text message links to financial structure applications; the process of existing recommendation strategies is not a closed-loop process, that is, the recommendation results cannot be used as new inputs to the process for the next round of recommendations; existing recommendation strategies cannot cope with high concurrency demands in large scenarios, and are prone to problems such as untimely responses or system crashes. Due to many limitations, the existing financial product recommendation strategies of financial institutions are ineffective and cannot attract enough users to browse or purchase financial products, which means that the yield of financial institutions cannot be effectively improved.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a method and device for recommending financial products, an electronic device and a program product, so as to at least solve the technical problem in the related art that the promotion efficiency of the configured financial products by financial institutions is low, resulting in low yield of the financial products.

[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for recommending financial products is provided, comprising: obtaining user feature data of a target user and corresponding financial product feature data from a distributed data warehouse, dividing the preprocessed user feature data and financial product feature data into N portions of feature data according to a preset rule, wherein N is a positive integer greater than 1, and the proportion of the first portion of feature data is greater than the proportion of the remaining N-1 portions of feature data; processing the N portions of feature data N-1 times using a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, and outputting N-1 correct recommendation quantities, wherein the Mth time The processing steps include: inputting the 1st to Mth feature data into the collaborative filtering algorithm to obtain the Mth recommended product list; comparing the financial products browsed and / or purchased by the target user in the M+1th user feature data with the Mth recommended product list to obtain the Mth correct recommendation number, where M is a positive integer from 1 to N-1; determining the target recommendation algorithm from the two collaborative filtering algorithms based on the N-1 correct recommendation numbers; inputting N pieces of feature data into the target recommendation algorithm to obtain the target recommended product list, and using the message middleware to send recommendation information to the user terminal held by the target user, wherein the recommendation information includes: the target recommended product list.

[0007] Optionally, the step of obtaining user characteristic data of the target user and the corresponding financial product characteristic data, and preprocessing the user characteristic data and the financial product characteristic data includes: obtaining the user characteristic data of the target user, the user characteristic data including at least one of the following: the target user's monthly salary income, the browsed financial products, and the purchased financial products, wherein each financial product is represented by a corresponding product ID number; obtaining financial product characteristic data corresponding to the browsed financial products and the purchased financial products, wherein the financial product characteristic data includes at least one of the following: product type, investment type, risk level, and fee structure; encoding the product type, investment type, risk level, and fee structure.

[0008] Optionally, the processing steps of the user-based collaborative filtering algorithm include: obtaining all non-target users and corresponding user feature data from a distributed data warehouse, and calculating the user similarity value between the target user and the non-target user based on the monthly salary income, browsed and / or purchased financial products in the user feature data of the target user and the non-target user; determining a first similar user among the non-target users based on the user similarity value between the target user and all non-target users; obtaining all financial products and corresponding users with browsing and / or purchasing behavior from the distributed data warehouse, wherein each user is represented by a corresponding user identification ID number; for each financial product, determining a non-target user who belongs to both the corresponding user with browsing and / or purchasing behavior and the first similar user as a second similar user, and calculating the target user's interest value in each financial product based on the user similarity value between the second similar user and the target user and the number of browsing and / or purchasing behaviors of the second similar user on the financial product; and determining a recommended product list based on the target user's interest value in all financial products.

[0009] Optionally, the processing steps of the product-based collaborative filtering algorithm include: defining the financial products browsed and / or purchased by the target user as target products; obtaining all financial product feature data from a distributed data warehouse, and calculating the product similarity value between each financial product and all other financial products except the financial product based on the financial product feature data; determining similar products of each financial product based on the product similarity values, and determining similar products belonging to the target product as target similar products; for each financial product, calculating the target user's interest value in the financial product based on the similarity value between the financial product and the corresponding target similar product and the number of browsing and / or purchasing behaviors of the target user on the target similar product; and determining a recommended product list based on the target user's interest values ​​in all financial products.

[0010] Optionally, the step of comparing the financial products browsed and / or purchased by the target user in the M+1th user characteristic data with the Mth recommended product list to obtain the Mth correct recommendation number includes: when the financial product browsed and / or purchased by the target user in the M+1th user characteristic data exists in the Mth recommended product list, determining the financial product as the correctly recommended product; and defining the number of correctly recommended products as the Mth correct recommendation number.

[0011] Optionally, the step of determining a target recommendation algorithm from two collaborative filtering algorithms based on the number of N-1 correct recommendations includes: calculating the accuracy of the two collaborative filtering algorithms based on the number of N-1 correct recommendations and the number of processing times of the collaborative filtering algorithms; sorting the accuracy rates, and determining the collaborative filtering algorithm with the highest accuracy in the sorting results as the target recommendation algorithm.

[0012] Optionally, after using the message middleware to send the recommendation information to the user terminal held by the target user, it also includes: receiving the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list sent by the user terminal; based on a preset timed task, using the database log message middleware to synchronize the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list to the relational database, wherein the relational database synchronizes the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list to the distributed data warehouse through the database log message middleware.

[0013] According to another aspect of an embodiment of the present invention, a financial product recommendation device is provided, comprising: an acquisition unit, configured to acquire user feature data of a target user and corresponding financial product feature data from a distributed data warehouse, and divide the pre-processed user feature data and financial product feature data into N portions of feature data according to a preset rule, wherein N is a positive integer greater than 1, and the proportion of the first portion of feature data is greater than the proportion of the remaining N-1 portions of feature data; a processing unit, configured to process the N portions of feature data N-1 times based on a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, and output N-1 correct recommendation quantities, wherein the Mth processing The steps include: inputting the 1st to Mth feature data into the collaborative filtering algorithm to obtain the Mth recommended product list; comparing the financial products browsed and / or purchased by the target user in the M+1th user feature data with the Mth recommended product list to obtain the Mth correct recommendation number, where M is a positive integer from 1 to N-1; a determination unit, which determines the target recommendation algorithm from the two collaborative filtering algorithms based on the N-1 correct recommendation numbers; a sending unit, which is used to input N pieces of feature data into the target recommendation algorithm to obtain the target recommended product list, and use the message middleware to send recommendation information to the user terminal held by the target user, wherein the recommendation information includes: the target recommended product list.

[0014] Optionally, the acquisition unit includes: a first acquisition module, used to acquire user characteristic data of the target user, the user characteristic data including at least one of the following: the target user's monthly salary income, browsed financial products, and purchased financial products, wherein each financial product is represented by a corresponding product ID number; a second acquisition module, used to acquire financial product characteristic data corresponding to the browsed financial products and purchased financial products, wherein the financial product characteristic data includes at least one of the following: product type, investment type, risk level, and fee structure; an encoding module, used to encode the product type, investment type, risk level, and fee structure.

[0015] Optionally, the processing unit includes: a first calculation module, which is used to obtain all non-target users and corresponding user feature data from a distributed data warehouse when executing a user-based collaborative filtering algorithm, and calculate the user similarity value between the target user and the non-target user based on the monthly salary income, browsed and / or purchased financial products in the user feature data of the target user and the non-target user; a first determination module, which determines a first similar user among the non-target users based on the user similarity value between the target user and all non-target users; a third acquisition module, which is used to obtain all financial products and corresponding users with browsing and / or purchasing behaviors from the distributed data warehouse, wherein each user is represented by a corresponding user identification ID number; a second calculation module, which, for each financial product, determines a non-target user who belongs to both the corresponding user with browsing and / or purchasing behaviors and the first similar user as a second similar user, and calculates the interest value of the target user in each financial product based on the user similarity value between the second similar user and the target user and the number of browsing and / or purchasing behaviors of the second similar user on the financial product; and a second determination module, which determines a recommended product list based on the interest value of the target user in all financial products.

[0016] Optionally, the processing unit also includes: a definition module, which is used to define the financial products browsed and / or purchased by the target user as target products when executing the product-based collaborative filtering algorithm; a third calculation module, which is used to obtain all financial product feature data from the distributed data warehouse, and calculate the product similarity value between each financial product and all other financial products except the financial product based on the financial product feature data; a third determination module, which determines the similar products of each financial product based on the product similarity value, and determines the similar products belonging to the target product as target similar products; a fourth calculation module, which calculates the interest value of the target user in the financial product for each financial product based on the similarity value between the financial product and the corresponding target similar product and the number of browsing and / or purchasing behaviors of the target user for the target similar product; and a fourth determination module, which determines the recommended product list based on the interest value of the target user in all financial products.

[0017] Optionally, the processing unit also includes: a fifth determination module, used to determine the financial product as a correctly recommended product when the financial product browsed and / or purchased by the target user in the M+1th user characteristic data exists in the Mth recommended product list; and a definition module, used to define the number of correctly recommended products as the Mth correct recommendation number.

[0018] Optionally, the determination unit includes: a fifth calculation module, used to calculate the accuracy of the two collaborative filtering algorithms based on the N-1 correct recommendation numbers and the number of processing times of the collaborative filtering algorithms when determining the target recommendation algorithm from the two collaborative filtering algorithms; a sixth determination module, used to sort the accuracy rates and determine the collaborative filtering algorithm with the highest accuracy in the sorting results as the target recommendation algorithm.

[0019] Optionally, the sending unit includes: a receiving module, which is used to receive the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list sent by the user terminal after sending the recommendation information to the user terminal held by the target user using the message middleware; a synchronization module, which uses the database log message middleware to synchronize the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list to the relational database based on a preset timed task, wherein the relational database synchronizes the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list to the distributed data warehouse through the database log message middleware.

[0020] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned methods for recommending financial products.

[0021] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned methods for recommending financial products.

[0022] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, the steps of any one of the above-mentioned methods for recommending financial products are implemented.

[0023] In the present disclosure, user feature data of a target user and corresponding financial product feature data are obtained from a distributed data warehouse, and the pre-processed user feature data and financial product feature data are divided into N pieces of feature data according to preset rules. The N pieces of feature data are then processed N-1 times by a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, and N-1 correct recommendation numbers are output. Then, a target recommendation algorithm is determined from the two collaborative filtering algorithms based on the N-1 correct recommendation numbers. Finally, the N pieces of feature data are input into the target recommendation algorithm to obtain a target recommended product list, and a message middleware is used to send recommendation information to a user terminal held by the target user, wherein the recommendation information includes: a target recommended product list.

[0024] According to the above disclosed content, by adopting two collaborative filtering algorithms, namely a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, the user feature data and the financial product feature data are processed multiple times, and the algorithm with better recommendation effect is selected from the two collaborative filtering algorithms to recommend financial products to the target users, thereby effectively improving the accuracy of financial product recommendations, enhancing the promotion efficiency of financial products, and increasing the rate of return of financial products, thereby solving the technical problem in the related technology that the promotion efficiency of financial institutions for the configured financial products is low, resulting in low rate of return of financial products. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0026] Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for recommending financial products is shown;

[0027] Figure 2 is a flow chart of an optional method for recommending financial products according to an embodiment of the present invention;

[0028] Figure 3 is a structural diagram of an optional financial product recommendation system according to an embodiment of the present invention;

[0029] Figure 4 is a flow chart of an optional adaptive collaborative filtering algorithm according to an embodiment of the present invention;

[0030] Figure 5 is a schematic diagram of an optional division of feature data according to an embodiment of the present invention;

[0031] Figure 6 is a schematic diagram of an optional financial product recommendation device according to an embodiment of the present invention;

[0032] Figure 7 It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0034] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0035] It should be noted that the financial product recommendation method and device in the present disclosure can be used in the field of financial technology to recommend financial products to target users, and can also be used in any field other than the field of financial technology to recommend financial products to target users. The present disclosure does not limit the application field of the financial product recommendation and authentication method and device.

[0036] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) collected by this disclosure are information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with the relevant laws, regulations and standards of the relevant regions, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation portals are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or organizations. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.

[0037] It should be noted that in the present disclosure, when collecting and analyzing customer information, corresponding operation entrances are provided for users to choose to agree or reject the automated decision results; if the user chooses to reject, the expert decision-making process is entered.

[0038] The following embodiments of the present invention can be applied to systems / applications / devices for recommending various financial products. The present invention obtains user feature data of target users and corresponding financial product feature data from a distributed data warehouse, divides the pre-processed user feature data and financial product feature data into N sets of feature data according to preset rules, and then processes the N sets of feature data N-1 times through a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, outputs N-1 correct recommendations, and then selects an algorithm with better recommendation effect from the two collaborative filtering algorithms based on the N-1 correct recommendations to recommend financial products to target users, thereby effectively improving the accuracy of financial product recommendations, enhancing the promotion efficiency of financial products, and thereby improving the yield of financial products.

[0039] The present invention is described in detail below in conjunction with various embodiments.

[0040] Embodiment 1

[0041] According to an embodiment of the present invention, an embodiment of a method for recommending financial products is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0042] The financial product recommendation method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for recommending financial products. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more ( Figure 1The computer system 102 includes a processor 102 (the processor 102 may include but is not limited to a microprocessor MCU (Microcontroller Unit) or a programmable logic device FPGA (Field Programmable Gate Array) and other processing devices), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0043] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0044] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the method for recommending financial products in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the method for recommending financial products mentioned above. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0045] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0046] The display may be, for example, a touch screen liquid crystal display (LCD), which enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0047] Under the above operating environment, this application provides Figure 2 Recommended methods for the financial products shown. Figure 2 is a flow chart of an optional method for recommending financial products according to an embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:

[0048] First of all, it should be noted that the embodiment of the present invention implements the recommendation method based on the recommendation system of financial products. Figure 3 is a structural diagram of an optional financial product recommendation system according to an embodiment of the present invention, such as Figure 3 As shown, the system includes a client system (corresponding to Figure 3 The "C-end system" in the middle) and the financial institution system (corresponding Figure 3 In the "B-end system" of the financial institution system, the financial institution creates financial products and stores the characteristic data of the financial products in the relational database. At the same time, the financial institution also stores the user characteristic data of the payroll users of the institution in the relational database.

[0049] In addition, the relational database synchronizes the financial product feature data and the user feature data to the distributed data warehouse through the database log message middleware. The user feature data and the financial product feature data stored in the distributed data warehouse are the basic data of the recommendation method of the embodiment of the present invention. The embodiment of the present invention generates recommended financial products for each user by obtaining these feature data from the distributed data warehouse.

[0050] Figure 3 The other contents in will be combined Figure 2 The following steps are illustrated.

[0051] Step S201, obtaining the user characteristic data of the target user and the corresponding financial product characteristic data from the distributed data warehouse, and dividing the pre-processed user characteristic data and financial product characteristic data into N sets of characteristic data according to a preset rule. Wherein N is a positive integer greater than 1, and the proportion of the first set of characteristic data is greater than the proportion of the remaining N-1 sets of characteristic data.

[0052] The purpose of the embodiment of the present invention is to make accurate financial product recommendations to target users. During the execution of the recommendation method, user feature data of the target user and corresponding financial product feature data are first obtained from a distributed data warehouse, and the two feature data are preprocessed.

[0053] It should be noted that if the target user has not browsed and / or purchased any financial products before, then when recommending to the target user, the obtained financial product feature data is the feature data of some financial products randomly selected from the distributed warehouse. If the target user has browsed and / or purchased any financial products before, the obtained financial product feature data is the feature data of the financial products browsed and / or purchased by the target user, that is, the "corresponding financial product feature data".

[0054] After preprocessing the user feature data and the financial product feature data, these feature data are input into the adaptive collaborative filtering algorithm provided by the embodiment of the present invention. The adaptive collaborative filtering algorithm combines the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm. Figure 4 is a flow chart of an optional adaptive collaborative filtering algorithm according to an embodiment of the present invention. Figure 4 As shown in the figure, in the "raw data preparation" stage, user feature data and financial product feature data are obtained from the distributed data warehouse. When entering the "data preparation" stage, the feature data is preprocessed. The preprocessing includes data cleaning, outlier detection, missing value filling and other steps to ensure the quality and consistency of the feature data.

[0055] After preprocessing the feature data, the feature data is divided into N portions of feature data according to a preset rule, wherein N is a positive integer greater than 1, and the proportion of the first portion of feature data is greater than the proportion of the remaining N-1 portions of feature data. Figure 5 is a schematic diagram of an optional division of feature data according to an embodiment of the present invention, such as Figure 5 As shown, the embodiment of the present invention can divide the feature data into four sets of feature data, A, B, C, and D, according to the order in which the target user browses and / or purchases the financial product, wherein set A accounts for 70%, and sets B, C, and D account for 10% respectively.

[0056] The purpose of dividing the feature data is to determine the more effective algorithm from the two collaborative filtering algorithms, namely the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm, so as to improve the accuracy of financial product recommendations.

[0057] Step S202, the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm process N pieces of feature data N-1 times, and output N-1 correct recommendation quantities, wherein the Mth processing step includes: inputting the 1st to Mth feature data into the collaborative filtering algorithm to obtain the Mth recommended product list; comparing the financial products browsed and / or purchased by the target user in the M+1th user feature data with the Mth recommended product list to obtain the Mth correct recommendation quantity, where M is a positive integer from 1 to N-1.

[0058] like Figure 4 As shown in the figure, after preprocessing the feature data and dividing it into N parts, the "pre-recommendation" stage is entered, and the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm are used to process the N parts of feature data N-1 times, and output N-1 correct recommendations. This process aims to compare the staged recommendations with the actual target user behavior, evaluate the performance of the two collaborative filtering algorithms on different target user feature data sets, and then select a more accurate and effective recommendation algorithm for the target user this time.

[0059] The process of the Mth processing is as follows: the first to the Mth feature data are input into the collaborative filtering algorithm, and then the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm are respectively executed. The user-based collaborative filtering algorithm will find user groups with similar behaviors and preferences as the target user, and use the browsing and / or purchase records of these groups to predict the financial products that the target user may be interested in, and then recommend them to the target user; the product-based collaborative filtering algorithm focuses on analyzing the similarities between financial products and recommending financial products that match the financial products that the target user has browsed and / or purchased. This process will output the Mth recommended product lists of the two collaborative filtering algorithms respectively, that is, for the target user, the set of financial products that the collaborative filtering algorithm predicts may be of interest. Subsequently, the financial products actually browsed and / or purchased by the target user in the M+1th user feature data are compared with the Mth recommended product list to determine the accuracy of the recommendation. Specifically, by checking how many financial products in the Mth recommended product list are actually browsed and / or purchased by the target user in the M+1th feature data, this number is the Mth correct recommendation number.

[0060] The above process is repeated N-1 times, each time processing a different feature data set, and eventually N-1 correct recommendations are obtained from the output of each of the two collaborative filtering algorithms.

[0061] Step S203: determining a target recommendation algorithm from two collaborative filtering algorithms based on the N-1 correct recommendation numbers.

[0062] like Figure 4 As shown, after using two collaborative filtering algorithms, namely, the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm, to process N pieces of feature data N-1 times and output N-1 correct recommendations, the "determine the target algorithm" stage is entered. Based on the output N-1 correct recommendations, a recommendation algorithm with higher accuracy and more in line with the target user's behavior pattern is determined from the two collaborative filtering algorithms as the target recommendation algorithm.

[0063] Step S204, inputting N pieces of feature data into a target recommendation algorithm to obtain a target recommended product list, and using a message middleware to send recommendation information to a user terminal held by a target user, wherein the recommendation information includes: a target recommended product list.

[0064] like Figure 4 As shown, after the target recommendation algorithm is determined, the "actual recommendation" stage is entered, and the full feature data of the target user (ie, N sets of feature data) is input into the target recommendation algorithm to obtain a target recommended product list.

[0065] The embodiment of the present invention can dynamically adjust the recommendation strategy according to different target users through multiple processing and algorithm comparison of two collaborative filtering algorithms, ensuring that when processing the characteristic data of the target users, the historical behavior patterns of the target users are taken into account, and the correlation between financial products is fully considered, ultimately achieving accurate and effective financial product recommendations.

[0066] After obtaining the target recommended product list, the embodiment of the present invention uses the message middleware to send the recommendation information to the user terminal held by the target user, wherein the recommendation information includes the target recommended product list. The recommendation information can be sent in the form of a text message, which includes a jump link to the target recommended product list for the target user to click and jump to enter the application of the financial institution to browse and / or purchase financial products.

[0067] like Figure 3 As shown, the embodiment of the present invention obtains feature data from a distributed data warehouse, obtains a target recommended product list by executing two collaborative filtering algorithms, and then uses a message middleware to send recommendation information to the target user in the form of a text message, that is, recommends financial products to the target user. After receiving the recommendation text message, the target user enters the application interface of the financial institution to browse the list and details of financial products, and purchase financial products.

[0068] Optionally, the step of obtaining user characteristic data of the target user and the corresponding financial product characteristic data, and preprocessing the user characteristic data and the financial product characteristic data includes: obtaining the user characteristic data of the target user, the user characteristic data including at least one of the following: the target user's monthly salary income, the browsed financial products, and the purchased financial products, wherein each financial product is represented by a corresponding product ID number; obtaining financial product characteristic data corresponding to the browsed financial products and the purchased financial products, wherein the financial product characteristic data includes at least one of the following: product type, investment type, risk level, and fee structure; encoding the product type, investment type, risk level, and fee structure.

[0069] In the embodiment of the present invention, since the adaptive collaborative filtering algorithm combines the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm, the feature data required by these two collaborative filtering algorithms are considered in the raw data preparation stage. The user feature data mainly includes: the monthly salary income of the target user, the browsed financial products, and the purchased financial products. The financial product feature data mainly includes: the financial product feature data corresponding to the financial products browsed and / or purchased by the target user, wherein the financial product feature data includes: product type, investment type, risk level, and fee structure.

[0070] After obtaining the above feature data, the non-numeric feature data is encoded, such as product type, investment type, risk level and fee structure. A numerical code is designed for each category based on their classification for subsequent input into the collaborative filtering algorithm for processing.

[0071] Optionally, the processing steps of the user-based collaborative filtering algorithm include: obtaining all non-target users and corresponding user feature data from a distributed data warehouse, and calculating the user similarity value between the target user and the non-target user based on the monthly salary income, browsed and / or purchased financial products in the user feature data of the target user and the non-target user; determining a first similar user among the non-target users based on the user similarity value between the target user and all non-target users; obtaining all financial products and corresponding users with browsing and / or purchasing behavior from the distributed data warehouse, wherein each user is represented by a corresponding user identification ID number; for each financial product, determining a non-target user who belongs to both the corresponding user with browsing and / or purchasing behavior and the first similar user as a second similar user, and calculating the target user's interest value in each financial product based on the user similarity value between the second similar user and the target user and the number of browsing and / or purchasing behaviors of the second similar user on the financial product; and determining a recommended product list based on the target user's interest value in all financial products.

[0072] The user-based collaborative filtering algorithm can be divided into two steps. The first step is to find a user cluster similar to the target user, and the second step is to find the financial products that the users in this cluster have browsed and / or purchased and recommend them to the target user.

[0073] During the execution of the user-based collaborative filtering algorithm, all non-target users and their corresponding user feature data are first obtained from the distributed data warehouse, and the user similarity values ​​between target users and non-target users are calculated based on their monthly salary income, browsed and / or purchased financial products in their user feature data.

[0074] It should be noted that users are stored in the distributed data warehouse in the form of ID numbers, and non-target users refer to all users in the distributed data warehouse except the target users.

[0075] When the target user has not browsed and / or purchased any financial products, the user similarity value between the target user and the non-target user is calculated based on the monthly salary income in the user feature data of the target user and the non-target user. The calculation formula is:

[0076]

[0077] In the above formula, x u ,y u 、z u are the monthly salary income of target user u in the past three months, x v ,y v 、z v They are the monthly salary income of non-target user v in the last three months.

[0078] When the target user has browsed and / or purchased financial products before, the user similarity value w between the target user and the non-target user is calculated based on the financial products browsed and / or purchased in the user feature data of the target user and the non-target user. uv , the calculation formula is:

[0079]

[0080] In the above formula, N(u) is the set of financial products browsed and / or purchased by target user u, and N(v) is the set of financial products browsed and / or purchased by non-target user v.

[0081] After calculating the user similarity values ​​between the target user and all non-target users, some non-target users with the largest user similarity values ​​are determined as first similar users of the target user.

[0082] Then, all financial products and corresponding users with browsing and / or purchasing behaviors are obtained from the distributed data warehouse. For each financial product, non-target users who belong to both the corresponding users with browsing and / or purchasing behaviors and the first category users are determined as second category users. That is, for financial product i, there is a user cluster with browsing and / or purchasing behaviors for it, and the user group among non-target users who belong to this user cluster and the first category users are determined as second category users.

[0083] Based on the user similarity value between the second similar user and the target user, and the number of times the second similar user browses and / or purchases the financial product, the interest value of the target user in each financial product is calculated. The calculation formula is:

[0084] p(u,i)=∑ v∈S(u，K)∩N(i) w uv r vi ,

[0085] In the above formula, p(u, i) is the interest value of target user u in financial product i, S(u, K) is the first similar user cluster (i.e., the K users most similar to target user u), N(i) is the user cluster that has browsed and / or purchased financial product i (i.e., the corresponding user cluster with browse and / or purchase behavior), v∈S(u, K)∩N(i) represents the second similar user v that belongs to both the corresponding user with browse and / or purchase behavior and the first similar user, w uv is the user similarity value between the target user u and the second similar user v, r vi The number of browsing and / or purchasing behaviors of the second type of user v on the financial product.

[0086] After calculating the interest values ​​of the target user for all financial products, H financial products are determined from high to low, and these H financial products are combined into a recommended product list, which is sent to the target user. The size of H is set by the user.

[0087] Optionally, the processing steps of the product-based collaborative filtering algorithm include: defining the financial products browsed and / or purchased by the target user as target products; obtaining all financial product feature data from a distributed data warehouse, and calculating the product similarity value between each financial product and all other financial products except the financial product based on the financial product feature data; determining similar products of each financial product based on the product similarity values, and determining similar products belonging to the target product as target similar products; for each financial product, calculating the target user's interest value in the financial product based on the similarity value between the financial product and the corresponding target similar product and the number of browsing and / or purchasing behaviors of the target user on the target similar product; and determining a recommended product list based on the target user's interest values ​​in all financial products.

[0088] The product-based collaborative filtering algorithm can be divided into two steps. The first step is to calculate the similarity values ​​between financial products, and the second step is to use the similarity values ​​between financial products to calculate the interest value of the target user in a certain financial product.

[0089] In the process of executing the product-based collaborative filtering algorithm, the financial products browsed and / or purchased by the target user are first defined as target products. It should be noted that the products are stored in the distributed data warehouse in the form of ID numbers.

[0090] Then, all financial product feature data are obtained from the distributed data warehouse, and the product similarity value between each financial product and all other financial products except the financial product is calculated based on the financial product feature data. Before calculating the product similarity value, the financial product feature data is first normalized, and the processing formula is:

[0091]

[0092] In the above formula, x norm is the normalized characteristic data of the f-th financial product, x is the original value of the characteristic data of the f-th financial product, max(x) is the maximum value in the characteristic data set of the f-th financial product, and min(x) is the minimum value in the characteristic data set of the f-th financial product.

[0093] Then calculate the product similarity value between the financial product and other financial products. The calculation formula is:

[0094]

[0095] In the above formula, sim(A, B) is the product similarity value between financial product A and other financial product B. is the characteristic data of the ith financial product of financial product A, is the characteristic data of the ith financial product of other financial product B, and n is the number of characteristic data of financial products.

[0096] After calculating the product similarity value between the financial product and all other financial products, select the largest part of other financial products as the same category products of the financial product. Then determine the same category products of the target product as the target same category products of the financial product. For example, for financial product A, by calculating the product similarity value between financial product A and all other financial products, determine the same category products of financial product A from other financial products. If there are financial products that the target user has browsed and / or purchased in the same category products, then these same category products are determined as the target same category products of financial product A. If there are no financial products that the target user has browsed and / or purchased in the same category products, then there is no target same category product for financial product A.

[0097] For each financial product, the target user's interest value for the financial product is calculated based on the similarity value between the financial product and the corresponding target similar product and the number of times the target user browses and / or purchases the target similar product. The calculation formula for the interest value is:

[0098]

[0099] In the above formula, p(u, j) is the interest value of target user u in financial product j, S(j, K) is the similar products of financial product j (i.e., the K financial products most similar to financial product j), N(u) is the set of financial products browsed and / or purchased by target user u (i.e., the target product set), i∈S(j, K)∩N(u) represents financial product i that is both a similar product and a target product (i.e., the target similar product i), W ji is the product similarity value between target similar product i and financial product j, R ui The number of times target user u browses and / or purchases target product i.

[0100] After calculating the interest values ​​of the target user for all financial products, H financial products are determined from high to low, and these H financial products are combined into a recommended product list, which is sent to the target user. The size of H is set by the user.

[0101] Optionally, the step of comparing the financial products browsed and / or purchased by the target user in the M+1th user characteristic data with the Mth recommended product list to obtain the Mth correct recommendation number includes: when the financial product browsed and / or purchased by the target user in the M+1th user characteristic data exists in the Mth recommended product list, determining the financial product as the correctly recommended product; and defining the number of correctly recommended products as the Mth correct recommendation number.

[0102] In step S202, the financial products browsed and / or purchased by the target user in the M+1th user characteristic data are compared with the Mth recommended product list to obtain the Mth correct recommendation number. The process is as follows: when the financial product browsed and / or purchased by the target user in the M+1th user characteristic data exists in the Mth recommended product list, the financial product is determined as the correct recommended product, and the number of correct recommended products is defined as the Mth correct recommendation number.

[0103] by Figure 5 Taking the partition feature data shown in the figure as an example, since the number of partitions N is 4, the two collaborative filtering algorithms are executed 3 times, as shown in the following Table 1:

[0104] Table 1

[0105]

[0106] In the first execution, the known data is A feature data, the known data to be predicted is B feature data, and the recommended product list output by the user-based collaborative filtering algorithm is B1. When the feature data to be predicted B∈B1, it means that the feature data is correctly recommended. That is, if there are 5 financial products in feature data B, 3 of which are in the recommended product list B1, then these 3 financial products are correctly recommended, and the remaining 2 financial products are not correctly recommended. The number of correctly recommended products "3" is defined as the Mth correct recommendation number. Similarly, the recommended product list output by the product-based collaborative filtering algorithm is B2. When the feature data to be predicted B∈B2, it means that the feature data is correctly recommended.

[0107] In the second execution, the known data is A+B feature data, the known data to be predicted is C feature data, the recommended product list output by the user-based collaborative filtering algorithm is C1, and when the feature data to be predicted C∈C1, it means that the feature data is correctly recommended; the recommended product list output by the product-based collaborative filtering algorithm is C2, and when the feature data to be predicted C∈C2, it means that the feature data is correctly recommended.

[0108] In the third execution, the known data is A+B+C feature data, the known data to be predicted is D feature data, the recommended product list output by the user-based collaborative filtering algorithm is D1, when the feature data to be predicted is D∈D1, it means that the feature data is correctly recommended; the recommended product list output by the product-based collaborative filtering algorithm is D2, when the feature data to be predicted is D∈D2, it means that the feature data is correctly recommended.

[0109] Optionally, the step of determining a target recommendation algorithm from two collaborative filtering algorithms based on the number of N-1 correct recommendations includes: calculating the accuracy of the two collaborative filtering algorithms based on the number of N-1 correct recommendations and the number of processing times of the collaborative filtering algorithms; sorting the accuracy rates, and determining the collaborative filtering algorithm with the highest accuracy in the sorting results as the target recommendation algorithm.

[0110] In step S203, the process of determining the target recommendation algorithm from the two collaborative filtering algorithms based on the number of N-1 correct recommendations is as follows: based on the number of N-1 correct recommendations and the number of processing times of the collaborative filtering algorithms, the accuracy of the two collaborative filtering algorithms is calculated, the accuracy is then sorted, and the collaborative filtering algorithm with the highest accuracy in the sorted results is determined as the target recommendation algorithm.

[0111] The accuracy calculation formula is: precision = hit / n, where precision is the accuracy, hit is the sum of the correct recommendations, and n is the number of executions.

[0112] by Figure 5 Taking the divided feature data shown as an example, if the number of correct recommendations of the user-based collaborative filtering algorithm in the first, second and third executions are 2, 3, and 4 respectively, and the number of correct recommendations of the product-based collaborative filtering algorithm in the first, second and third executions are 3, 4, and 5 respectively, then the calculation formula of the user-based collaborative filtering algorithm is: precision = (2+3+4) / 3; the calculation formula of the product-based collaborative filtering algorithm is: precision = (3+4+5) / 3.

[0113] After calculating the accuracy of the two collaborative filtering algorithms, the accuracy is sorted, and the collaborative filtering algorithm with the highest accuracy in the sorted results is determined as the target recommendation algorithm.

[0114] Optionally, after using the message middleware to send the recommendation information to the user terminal held by the target user, it also includes: receiving the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list sent by the user terminal; based on a preset timed task, using the database log message middleware to synchronize the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list to the relational database, wherein the relational database synchronizes the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list to the distributed data warehouse through the database log message middleware.

[0115] It should be noted that the embodiments of the present invention take into account the performance issues under high concurrency, that is, browsing and / or purchase requests initiated by a large number of target users, and store the financial product feature data in the cache middleware. At the same time, the target users' behavior data on financial products, including browsing and / or purchase behavior data, is stored in the cache middleware to cope with high concurrency issues in large scenarios.

[0116] At the same time, if Figure 3 As shown, the embodiment of the present invention synchronizes the browsing and / or purchasing behavior data of the target user on the financial products in the target recommended product list stored in the cache middleware to the relational database based on the preset timed task. Financial institutions can create, query, and modify financial products through the relational database, and can also analyze the recommendation effect by calling the data stored in the relational database.

[0117] In addition, the relational database synchronizes the browsing and / or purchasing behavior data of the target user on the target recommended product list to the distributed data warehouse tool through the database log message middleware. These latest behavior data will be used in the next round of product recommendations for the target user, thus forming a closed-loop process structure. By recommending financial products to the target user, collecting the latest dynamics of the target user, and then making the next round of product recommendations to the target user based on the latest dynamics, the accuracy of financial product recommendations can be continuously improved, the promotion efficiency of financial products can be improved, and the yield of financial products can be increased.

[0118] It should be noted that Figure 3 The financial product recommendation system shown in the figure adopts the form of end isolation and hot-cold isolation. End isolation refers to the isolation between the C end and the B end. The C end uses the message middleware to reduce the peak and fill the valley of the traffic sent by the recommendation SMS, reduce the pressure on the system server and improve the utilization rate of the system server. At the same time, the C end uses the cache middleware to handle the high concurrent requests of the target users, thereby improving the high concurrent processing capability of the system. Since the B end has a small amount of access, it uses a relational database and a distributed data warehouse to store data and process business. Data is transmitted between the B end and the C end through preset timed tasks and database log message middleware, ensuring the data synchronization between the B end and the C end. Hot-cold isolation refers to deploying the message middleware on the C end to store feature data for frequent access by the target users, and deploying the relational database on the B end to store feature data for small-scale access by financial institutions. By isolating the frequent access of the target users from the small-scale access of the financial institutions, the overall performance of the system can be improved.

[0119] The steps provided by the above-mentioned financial product recommendation method are to obtain the user feature data of the target user and the corresponding financial product feature data from the distributed data warehouse, divide the pre-processed user feature data and the financial product feature data into N pieces of feature data according to preset rules, and then process the N pieces of feature data N-1 times through the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm, output N-1 correct recommendation numbers, and then determine the target recommendation algorithm from the two collaborative filtering algorithms based on the N-1 correct recommendation numbers, and finally input the N pieces of feature data into the target recommendation algorithm to obtain a target recommended product list, and use the message middleware to send recommendation information to the user terminal held by the target user, wherein the recommendation information includes: a target recommended product list.

[0120] The embodiment of the present invention adopts two collaborative filtering algorithms, namely a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, to process user feature data and financial product feature data multiple times, and selects an algorithm with better recommendation effect from the two collaborative filtering algorithms to recommend financial products to target users, thereby effectively improving the accuracy of financial product recommendations, enhancing the promotion efficiency of financial products, and at the same time improving the yield of financial products, thereby solving the technical problem in related technologies that financial institutions have low promotion efficiency for configured financial products, resulting in low yield of financial products.

[0121] The following is a detailed description in conjunction with another embodiment.

[0122] Embodiment 2

[0123] A financial product recommendation device provided in this embodiment includes multiple implementation units, each of which corresponds to each implementation step in the above-mentioned embodiment 1.

[0124] Figure 6 is a schematic diagram of an optional financial product recommendation device according to an embodiment of the present invention, such as Figure 6 As shown, the recommendation device based on financial products may include: an acquisition unit 61 , a processing unit 62 , a determination unit 63 , and a sending unit 64 .

[0125] Among them, the acquisition unit 61 is used to obtain the user feature data of the target user and the corresponding financial product feature data from the distributed data warehouse, and divide the pre-processed user feature data and financial product feature data into N sets of feature data according to preset rules, wherein N is a positive integer greater than 1, and the proportion of the first set of feature data is greater than the proportion of the remaining N-1 sets of feature data.

[0126] The processing unit 62 processes N pieces of feature data N-1 times based on a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, and outputs N-1 correct recommendation quantities, wherein the M-th processing step includes: inputting the 1st to M-th feature data into the collaborative filtering algorithm to obtain the M-th recommended product list; comparing the financial products browsed and / or purchased by the target user in the M+1-th user feature data with the M-th recommended product list to obtain the M-th correct recommendation quantity, where M is a positive integer from 1 to N-1.

[0127] The determination unit 63 determines a target recommendation algorithm from two collaborative filtering algorithms based on the N-1 correct recommendation numbers.

[0128] The sending unit 64 is used to input N pieces of feature data into the target recommendation algorithm to obtain a target recommended product list, and use the message middleware to send recommendation information to the user terminal held by the target user, wherein the recommendation information includes: the target recommended product list.

[0129] The above-mentioned financial product recommendation device can obtain the user feature data of the target user and the corresponding financial product feature data from the distributed data warehouse through the acquisition unit 61, divide the pre-processed user feature data and financial product feature data into N pieces of feature data according to preset rules, and then use the processing unit 62 to process the N pieces of feature data N-1 times based on the user's collaborative filtering algorithm and the product's collaborative filtering algorithm, output N-1 correct recommendation numbers, and then use the determination unit 63 to determine the target recommendation algorithm from the two collaborative filtering algorithms based on the N-1 correct recommendation numbers, and finally use the sending unit 64 to input the N pieces of feature data into the target recommendation algorithm to obtain a target recommended product list, and use the message middleware to send recommendation information to the user terminal held by the target user, wherein the recommendation information includes: a target recommended product list. The embodiment of the present invention adopts two collaborative filtering algorithms, namely a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, to process user feature data and financial product feature data multiple times, and selects an algorithm with better recommendation effect from the two collaborative filtering algorithms to recommend financial products to target users, thereby effectively improving the accuracy of financial product recommendations, enhancing the promotion efficiency of financial products, and increasing the rate of return of financial products, thereby solving the technical problem in related technologies that financial institutions have low promotion efficiency for configured financial products, resulting in low rate of return of financial products.

[0130] Optionally, the acquisition unit 61 includes: a first acquisition module, used to acquire user characteristic data of the target user, the user characteristic data including at least one of the following: the target user's monthly salary income, browsed financial products, and purchased financial products, wherein each financial product is represented by a corresponding product ID number; a second acquisition module, used to acquire financial product characteristic data corresponding to the browsed financial products and purchased financial products, wherein the financial product characteristic data includes at least one of the following: product type, investment type, risk level, and fee structure; an encoding module, used to encode the product type, investment type, risk level, and fee structure.

[0131] Optionally, the processing unit 62 includes: a first calculation module, which is used to obtain all non-target users and corresponding user feature data from the distributed data warehouse when executing the user-based collaborative filtering algorithm, and calculate the user similarity value between the target user and the non-target user based on the monthly salary income, browsed and / or purchased financial products in the user feature data of the target user and the non-target user; a first determination module, which determines the first similar user among the non-target users based on the user similarity value between the target user and all non-target users; a third acquisition module, which is used to obtain all financial products and corresponding users with browsing and / or purchasing behaviors from the distributed data warehouse, wherein each user is represented by a corresponding user identification ID number; a second calculation module, which, for each financial product, determines the non-target user who belongs to both the corresponding user with browsing and / or purchasing behaviors and the first similar user as a second similar user, and calculates the interest value of the target user in each financial product based on the user similarity value between the second similar user and the target user and the number of browsing and / or purchasing behaviors of the second similar user on the financial product; and a second determination module, which determines the recommended product list based on the interest value of the target user in all financial products.

[0132] Optionally, the processing unit 62 also includes: a definition module, which is used to define the financial products browsed and / or purchased by the target user as target products when executing the product-based collaborative filtering algorithm; a third calculation module, which is used to obtain all financial product feature data from the distributed data warehouse, and calculate the product similarity value between each financial product and all other financial products except the financial product based on the financial product feature data; a third determination module, which determines the similar products of each financial product based on the product similarity value, and determines the similar products belonging to the target product as target similar products; a fourth calculation module, which calculates the interest value of the target user in the financial product for each financial product based on the similarity value between the financial product and the corresponding target similar product and the number of browsing and / or purchasing behaviors of the target user on the target similar product; and a fourth determination module, which determines the recommended product list based on the interest value of the target user in all financial products.

[0133] Optionally, the processing unit 62 also includes: a fifth determination module, used to determine the financial product as a correct recommended product when the financial product browsed and / or purchased by the target user in the M+1th user characteristic data exists in the Mth recommended product list; and a definition module, used to define the number of correct recommended products as the Mth correct recommendation number.

[0134] Optionally, the determination unit 63 includes: a fifth calculation module, which is used to calculate the accuracy of the two collaborative filtering algorithms based on the N-1 correct recommendation numbers and the number of processing times of the collaborative filtering algorithms when determining the target recommendation algorithm from the two collaborative filtering algorithms; a sixth determination module, which is used to sort the accuracy rates and determine the collaborative filtering algorithm with the highest accuracy in the sorting results as the target recommendation algorithm.

[0135] Optionally, the sending unit 64 includes: a receiving module, which is used to receive the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list sent by the user terminal after the recommendation information is sent to the user terminal held by the target user using the message middleware; a synchronization module, which uses the database log message middleware to synchronize the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list to the relational database based on a preset timing task, wherein the relational database synchronizes the target user's browsing and / or purchasing behavior data on the financial products in the target recommended product list to the distributed data warehouse through the database log message middleware.

[0136] The financial product recommendation device may further include a processor and a memory. The acquisition unit 61, the processing unit 62, the determination unit 63, the sending unit 64, etc. are all stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions.

[0137] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and financial products can be recommended to target users by adjusting kernel parameters.

[0138] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip.

[0139] Embodiment 3

[0140] An embodiment of the present application may provide an electronic device, Figure 7 is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 7 As shown, the electronic device may include: one or more ( Figure 7 (only one is shown) processor 702, memory 704, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0141] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0142] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: obtain the user feature data of the target user and the corresponding financial product feature data from the distributed data warehouse, divide the pre-processed user feature data and financial product feature data into N pieces of feature data according to preset rules, wherein N is a positive integer greater than 1, and the proportion of the first piece of feature data is greater than the proportion of the remaining N-1 pieces of feature data; process the N pieces of feature data N-1 times based on the user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm, and output N-1 correct recommendation quantities, wherein the Mth processing The processing steps include: inputting the 1st to Mth feature data into the collaborative filtering algorithm to obtain the Mth recommended product list; comparing the financial products browsed and / or purchased by the target user in the M+1th user feature data with the Mth recommended product list to obtain the Mth correct recommendation number, where M is a positive integer from 1 to N-1; determining the target recommendation algorithm from the two collaborative filtering algorithms based on the N-1 correct recommendation numbers; inputting N pieces of feature data into the target recommendation algorithm to obtain the target recommended product list, and using the message middleware to send recommendation information to the user terminal held by the target user, wherein the recommendation information includes: the target recommended product list.

[0143] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: obtain user characteristic data of the target user, the user characteristic data including at least one of the following: the target user's monthly salary income, the browsed financial products, and the purchased financial products, wherein each financial product is represented by a corresponding product ID number; obtain financial product characteristic data corresponding to the browsed financial products and the purchased financial products, wherein the financial product characteristic data includes at least one of the following: product type, investment type, risk level, and fee structure; and encode the product type, investment type, risk level, and fee structure.

[0144] The processor can also call the information and application stored in the memory through the transmission device to perform the following steps: obtain all non-target users and corresponding user feature data from the distributed data warehouse, and calculate the user similarity value between the target user and the non-target user based on the monthly salary income, browsed and / or purchased financial products in the user feature data of the target user and the non-target user; determine the first similar user among the non-target users based on the user similarity value between the target user and all non-target users; obtain all financial products and corresponding users with browsing and / or purchasing behavior from the distributed data warehouse, wherein each user is represented by a corresponding user identification ID number; for each financial product, determine the non-target user who belongs to both the corresponding user with browsing and / or purchasing behavior and the first similar user as a second similar user, and calculate the interest value of the target user in each financial product based on the user similarity value between the second similar user and the target user and the number of browsing and / or purchasing behaviors of the second similar user on the financial product; determine a recommended product list based on the interest value of the target user in all financial products.

[0145] The processor can also call the information and application stored in the memory through the transmission device to perform the following steps: define the financial products browsed and / or purchased by the target user as target products; obtain all financial product feature data from the distributed data warehouse, and calculate the product similarity value between each financial product and all other financial products except the financial product based on the financial product feature data; determine the similar products of each financial product based on the product similarity value, and determine the similar products of the target product as target similar products; for each financial product, calculate the interest value of the target user in the financial product based on the similarity value between the financial product and the corresponding target similar product and the number of browsing and / or purchasing behaviors of the target user for the target similar product; and determine a recommended product list based on the interest value of the target user in all financial products.

[0146] The processor can also call the information and application stored in the memory through the transmission device to perform the following steps: when the financial product browsed and / or purchased by the target user in the M+1th user characteristic data exists in the Mth recommended product list, the financial product is determined as the correct recommended product; and the number of correctly recommended products is defined as the Mth correct recommendation number.

[0147] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: based on the number of N-1 correct recommendations and the number of processing times of the collaborative filtering algorithm, calculate the accuracy of the two collaborative filtering algorithms; sort the accuracy rates, and determine the collaborative filtering algorithm with the highest accuracy in the sorting results as the target recommendation algorithm.

[0148] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: receiving the browsing and / or purchasing behavior data of the target user on the financial products in the target recommended product list sent by the user end; based on the preset timing task, using the database log message middleware to synchronize the browsing and / or purchasing behavior data of the target user on the financial products in the target recommended product list to the relational database, wherein the relational database synchronizes the browsing and / or purchasing behavior data of the target user on the financial products in the target recommended product list to the distributed data warehouse through the database log message middleware.

[0149] By adopting the embodiment of the present application, a scheme of a method for recommending financial products is provided, by obtaining user feature data of a target user and corresponding financial product feature data from a distributed data warehouse, dividing the pre-processed user feature data and financial product feature data into N pieces of feature data according to preset rules, and then processing the N pieces of feature data N-1 times by a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, outputting N-1 correct recommendation numbers, and then determining a target recommendation algorithm from the two collaborative filtering algorithms based on the N-1 correct recommendation numbers, and finally obtaining a target recommended product list by inputting the N pieces of feature data into the target recommendation algorithm, and using a message middleware to send recommendation information to a user terminal held by the target user, wherein the recommendation information includes: a target recommended product list, thereby achieving the purpose of effectively improving the accuracy of financial product recommendations, enhancing the promotion efficiency of financial products, and increasing the rate of return of financial products, thereby solving the technical problem in the related technology that the promotion efficiency of financial institutions for configured financial products is low, resulting in low rate of return of financial products.

[0150] It can be understood by those skilled in the art that Figure 7The structure shown is for illustration only, and the electronic device may also be a terminal device such as a smart phone, a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, etc. Figure 7 The structure of the electronic device is not limited. Figure 7 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 7 Different configurations shown.

[0151] A person of ordinary skill in the art can understand that all or part of the steps in the recommendation method for various financial products in the above-mentioned embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0152] Embodiment 4

[0153] The embodiment of the present application further provides a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the financial product recommendation method provided in the first embodiment.

[0154] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the financial product recommendation method of any one of the above-mentioned embodiments one.

[0155] Optionally, in this embodiment, the above storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0156] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the financial product recommendation method described in each embodiment of the present application.

[0157] The present application also provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the financial product recommendation method described in each embodiment of the present application are implemented.

[0158] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0159] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0160] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0161] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0162] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0163] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0164] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for recommending financial products, characterized in that: include: Obtaining user characteristic data of a target user and corresponding financial product characteristic data from a distributed data warehouse, and dividing the preprocessed user characteristic data and the financial product characteristic data into N portions of characteristic data according to a preset rule, wherein N is a positive integer greater than 1, and the proportion of the first portion of characteristic data is greater than the proportion of the remaining N-1 portions of characteristic data; The user-based collaborative filtering algorithm and the product-based collaborative filtering algorithm process the N pieces of feature data N-1 times, and output N-1 correct recommendation quantities, wherein the step of the Mth processing comprises: inputting the 1st to Mth pieces of feature data into the collaborative filtering algorithm to obtain the Mth recommended product list; comparing the financial products browsed and / or purchased by the target user in the M+1th piece of user feature data with the Mth recommended product list to obtain the Mth correct recommendation quantity, where M is a positive integer from 1 to N-1; Determine a target recommendation algorithm from two collaborative filtering algorithms based on the N-1 correct recommendation numbers; The N pieces of feature data are input into the target recommendation algorithm to obtain a target recommended product list, and a message middleware is used to send recommendation information to a user terminal held by the target user, wherein the recommendation information includes: the target recommended product list.

2. The recommendation method according to claim 1, characterized in that: The step of obtaining user characteristic data of the target user and corresponding financial product characteristic data, and preprocessing the user characteristic data and the financial product characteristic data comprises: Acquire the user characteristic data of the target user, the user characteristic data including at least one of the following: the monthly salary income, the browsed financial products, and the purchased financial products of the target user, wherein each financial product is represented by a corresponding product ID number; Acquire financial product characteristic data corresponding to the browsed financial product and the purchased financial product, wherein the financial product characteristic data includes at least one of the following: product type, investment type, risk level, and fee structure; The product type, the investment type, the risk level, and the fee structure are coded.

3. The recommendation method according to claim 1, characterized in that: The processing steps of the user-based collaborative filtering algorithm include: Acquire all non-target users and corresponding user feature data from the distributed data warehouse, and calculate the user similarity value between the target user and the non-target user based on the monthly salary income, browsed and / or purchased financial products in the user feature data of the target user and the non-target user; Based on the user similarity values ​​between the target user and all the non-target users, determining a first similar user among the non-target users; Obtaining all financial products and corresponding users who have browsed and / or purchased the products from the distributed data warehouse, wherein each user is represented by a corresponding user identification ID number; For each financial product, a non-target user who is both a corresponding user with browsing and / or purchasing behavior and the first similar user is determined as a second similar user, and based on the user similarity value between the second similar user and the target user and the number of browsing and / or purchasing behaviors of the second similar user on the financial product, the interest value of the target user in each financial product is calculated; The recommended product list is determined based on the interest values ​​of the target user in all financial products.

4. The recommendation method according to claim 1, characterized in that: The processing steps of the product-based collaborative filtering algorithm include: defining the financial product browsed and / or purchased by the target user as a target product; Acquire all financial product feature data from the distributed data warehouse, and calculate the product similarity value between each financial product and all other financial products except the financial product based on the financial product feature data; Based on the product similarity value, determine similar products of each financial product, and determine the similar products belonging to the target product as target similar products; For each financial product, based on the similarity between the financial product and the corresponding target similar product and the number of browsing and / or purchasing behaviors of the target user on the target similar product, the interest value of the target user in the financial product is calculated; The recommended product list is determined based on the interest values ​​of the target user in all financial products.

5. The recommendation method according to claim 1, characterized in that: The step of comparing the financial products browsed and / or purchased by the target user in the M+1th user characteristic data with the Mth recommended product list to obtain the Mth correct recommended quantity includes: If the financial product browsed and / or purchased by the target user in the M+1th piece of user characteristic data exists in the Mth recommended product list, the financial product is determined as a correct recommended product; The number of the correctly recommended products is defined as the Mth correctly recommended number.

6. The recommendation method according to claim 1, characterized in that: The step of determining the target recommendation algorithm from the two collaborative filtering algorithms based on the N-1 correct recommendation numbers includes: Based on the number of N-1 correct recommendations and the number of times the collaborative filtering algorithm is processed, the accuracy of the two collaborative filtering algorithms is calculated; The accuracy rates are sorted, and the collaborative filtering algorithm with the highest accuracy rate in the sorted results is determined as the target recommendation algorithm.

7. The recommendation method according to claim 1, characterized in that: After the message middleware is used to send the recommendation information to the user terminal held by the target user, the method further includes: Receiving browsing and / or purchasing behavior data of the target user on the financial products in the target recommended product list sent by the user terminal; Based on the preset timed tasks, the database log message middleware is used to synchronize the target user's browsing and / or purchasing behavior data of the financial products in the target recommended product list to the relational database, wherein the relational database synchronizes the target user's browsing and / or purchasing behavior data of the financial products in the target recommended product list to the distributed data warehouse through the database log message middleware.

8. A financial product recommendation device, characterized in that: include: An acquisition unit is used to acquire user characteristic data of a target user and corresponding financial product characteristic data from a distributed data warehouse, and divide the pre-processed user characteristic data and the financial product characteristic data into N portions of characteristic data according to a preset rule, wherein N is a positive integer greater than 1, and a proportion of the first portion of characteristic data is greater than a proportion of the remaining N-1 portions of characteristic data; A processing unit processes the N pieces of feature data N-1 times based on a user-based collaborative filtering algorithm and a product-based collaborative filtering algorithm, and outputs N-1 correct recommendation quantities, wherein the step of the M-th processing comprises: inputting the 1st to M-th pieces of feature data into the collaborative filtering algorithm to obtain an M-th recommended product list; comparing the financial products browsed and / or purchased by the target user in the M+1th piece of user feature data with the M-th recommended product list to obtain an M-th correct recommendation quantity, where M is a positive integer from 1 to N-1; A determination unit, which determines a target recommendation algorithm from two collaborative filtering algorithms based on the N-1 correct recommendation numbers; The sending unit is used to input the N pieces of feature data into the target recommendation algorithm to obtain a target recommended product list, and use a message middleware to send recommendation information to a user terminal held by the target user, wherein the recommendation information includes: the target recommended product list.

9. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the financial product recommendation method described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method for recommending financial products according to any one of claims 1 to 7 are implemented.

Images