WO2020048051A1 - Financial product recommendation method, server and computer readable storage medium - Google Patents

Abstract

A financial product recommendation method, a server and a computer readable storage medium. Said method comprises: collecting a user profile of a user, the user profile comprising behavior data and user attributes (S400); performing risk capability classification on the user on the basis of a random forest model, so as to obtain a risk classification result of the user (S402); performing liquidity positioning on the user according to the user profile, so as to obtain a liquidity positioning result of the user (S404); and acquiring a first minimum investment amount of the user, and recommending, in conjunction with the risk classification result and the liquidity positioning result, an investment product suitable for the user (S406). Said method is able to combine the risk tolerance of a user and the demand for funding liquidity to provide the user with an optimum financial product, so as to maximize revenue.

Description

Financial product recommendation method, server and computer-readable storage medium

This application claims priority from a Chinese patent application filed with the Chinese Patent Office on September 4, 2018, with application number 201811024273.4, and the invention name is "Finance Product Recommendation Method, Server, and Computer-readable Storage Medium", the entire contents of which are incorporated by reference Incorporated in this application.

Technical field

The present application relates to the field of data analysis technology, and in particular, to a method for recommending wealth management products, a server, and a computer-readable storage medium.

Background technique

China's personal financial industry is still in a rapid development stage. Most users fill out questionnaires to determine their own risk tolerance, and then find corresponding financial products. Therefore, there are the following disadvantages: On the one hand, users may be affected by personal psychology and other factors when filling out the questionnaire, so the results of the questionnaire alone cannot accurately describe the user's risk tolerance, and there is great uncertainty Or inaccuracy; on the other hand, different users have different needs for liquidity (reflecting the time and scale characteristics of fund withdrawal, including: bullet type, step type, maturity, and dumbbell type, and capital), and dynamic returns, and dynamic balance The relationship between the two is difficult. Among the many wealth management products, from risk tolerance, liquidity to income, how to identify the most suitable product is more difficult for users without investment experience. At present, there are no similar products on the market. According to the user's asset ratio and investment path, the asset ratio planning is performed.

Summary of the Invention

In view of this, this application proposes a method for recommending wealth management products, a server, and a computer-readable storage medium to solve a user's selection problem when facing many wealth management products.

First, in order to achieve the above purpose, this application proposes a method for recommending wealth management products. The method includes steps:

Collect user data of users;

Classifying the user's risk ability based on a random forest model to obtain a risk classification result for the user;

Locating the user's mobility according to the user profile to obtain a result of the user's mobility positioning; and

Obtain the user's minimum investment amount, and combine the risk classification result and the liquidity positioning result to recommend investment products suitable for the user.

In addition, in order to achieve the above object, the present application further provides a server including a memory and a processor, where the memory stores a financial product recommendation system that can be run on the processor, and the financial product recommendation system is described by The processor executes the steps of the method for recommending a financial product as described above.

Further, in order to achieve the above object, the present application further provides a computer-readable storage medium, where the computer-readable storage medium stores a financial product recommendation system, and the financial product recommendation system can be executed by at least one processor so that The at least one processor executes the steps of the financial product recommendation method as described above.

Compared with the prior art, the financial product recommendation method, server, and computer-readable storage medium proposed in the present application can collect user behavior data and user attributes and classify the user's risk ability based on a random forest model to obtain The user's risk classification result, then the user's liquidity positioning is performed according to the user's historical purchase record, the user's liquidity positioning result is obtained, and the risk classification result is converted into a continuity score, and finally obtained The user's minimum investment, combined with the risk classification result and the liquidity positioning result, recommends investment products suitable for the user. This application combines the user's risk tolerance and the need for liquidity, and uses internationally recognized securities investment strategy theory and the industry's first liquidity preference model to divide users into different types of financial management based on their investment history. , And developed a set of exclusive financial planning recommendations for various types of users, this proposal based on the existing financial structure of the user to "fill the vacancies", so as to help users to achieve maximum win-win risk, liquidity and returns.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of an optional hardware architecture of a server of the present application;

2 is a schematic diagram of a program module of a first embodiment of a financial product recommendation system of the present application;

3 is a schematic diagram of a program module of a second embodiment of a financial product recommendation system of the present application;

4 is a schematic flowchart of a first embodiment of a method for recommending wealth management products of the present application;

5 is a schematic flowchart of a second embodiment of a method for recommending wealth management products of the present application;

6 is a schematic flowchart of a third embodiment of a method for recommending wealth management products of the present application;

FIG. 7 is a schematic flowchart of a fourth embodiment of a method for recommending wealth management products of the present application.

Reference signs:

server	2
Memory	11
processor	12
Network Interface	13
Financial product recommendation system	200
Collection module	201
Risk Classification Module	202
Liquidity positioning module	203

Recommended modules	204
Conversion module	205

The implementation, functional features and advantages of the purpose of this application will be further described with reference to the embodiments and the drawings.

detailed description

In order to make the purpose, technical solution, and advantages of the present application clearer, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the application, and are not used to limit the application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

It should be noted that the descriptions related to "first" and "second" in this application are only for descriptive purposes, and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated . Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but must be based on those that can be realized by a person of ordinary skill in the art. When the combination of technical solutions conflicts or cannot be achieved, it should be considered that such a combination of technical solutions does not exist. Is not within the scope of protection claimed in this application.

FIG. 1 is a schematic diagram of an optional hardware architecture of the server 2 of the present application.

In this embodiment, the server 2 may include, but is not limited to, a memory 11, a processor 12, and a network interface 13 which may communicate with each other through a system bus. It should be noted that FIG. 1 only shows the server 2 with components 11-13, but it should be understood that it is not required to implement all the illustrated components, and more or fewer components may be implemented instead.

The server 2 may be a computing device such as a rack server, a blade server, a tower server, or a rack server. The server 2 may be an independent server or a server cluster composed of multiple servers.

The memory 11 includes at least one type of readable storage medium. The readable storage medium includes a flash memory, a hard disk, a multimedia card, a card-type memory (for example, SD or DX memory, etc.), a random access memory (RAM), and a static memory. Random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disks, optical disks, etc. In some embodiments, the memory 11 may be an internal storage unit of the server 2, such as a hard disk or a memory of the server 2. In other embodiments, the memory 11 may also be an external storage device of the server 2, such as a plug-in hard disk, a Smart Memory Card (SMC), and a secure digital (Secure) Digital, SD) cards, flash cards, etc. Of course, the memory 11 may also include both an internal storage unit of the server 2 and an external storage device thereof. In this embodiment, the memory 11 is generally used to store an operating system and various application software installed on the server 2, such as program codes of the financial product recommendation system 200. In addition, the memory 11 may also be used to temporarily store various types of data that have been output or will be output.

In some embodiments, the processor 12 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or another data processing chip. The processor 12 is generally used to control the overall operation of the server 2. In this embodiment, the processor 12 is configured to run program code or process data stored in the memory 11, for example, to run the financial product recommendation system 200 and the like.

The network interface 13 may include a wireless network interface or a wired network interface. The network interface 13 is generally used to establish a communication connection between the server 2 and other electronic devices.

So far, the hardware structure and functions of the related equipment of this application have been introduced in detail. In the following, various embodiments of the present application will be made based on the above description.

First, this application proposes a financial product recommendation system 200.

Referring to FIG. 2, it is a program module diagram of the first embodiment of the financial product recommendation system 200 of the present application.

In this embodiment, the financial product recommendation system 200 includes a series of computer program instructions stored in the memory 11. When the computer program instructions are executed by the processor 12, the financial product recommendation operations of the embodiments of the present application can be implemented . In some embodiments, the financial product recommendation system 200 may be divided into one or more modules based on specific operations implemented by various portions of the computer program instructions. For example, in FIG. 2, the financial product recommendation system 200 may be divided into a collection module 201, a risk classification module 202, a liquidity positioning module 203, and a recommendation module 204. among them:

The collection module 201 is configured to collect user data of a user, where the user data includes behavior data and user attributes.

In one embodiment, the collection module 201 may obtain the user profile from application software. The application software may be an account communication APP.

Specifically, the user attributes include gender, age, consumption level, income level, investment experience, and registration days.

The behavior data includes a user's operation record on the mobile terminal, and specifically includes a user's operation record on the One Account App, and the operation record may include a click operation, a browsing operation, and a consumption record.

The risk classification module 202 is configured to establish a sample set according to valid user data, classify the user's risk ability based on a random forest model, and obtain a risk classification result of the user.

Specifically, the risk classification result may include a conservative type, a stable type, a balanced type, a growth type, and an aggressive type.

The effective user information is collected user information of multiple groups of users and corresponding risk classification results, wherein the user information includes gender, age, consumption level, income level, investment experience, registration days, click operation, and browse Features such as operations, consumption records are used as independent variables of the sample set, and the risk classification results are used as dependent variables of the sample set.

Further, the step of classifying the risk capability of the user by the risk classification module 202 to obtain the risk classification result of the user includes:

Random forest model training process:

(a) establish a sample set based on valid user data;

(b) Generate n feature samples from the sample set by resampling. The number of feature components in each feature sample is a, the number of decision trees is m, and the number of decision features of each decision tree. K

(c) Using Bagging algorithm (bagging algorithm) to sample n feature samples m times to obtain m feature sets;

(d) Randomly select a feature set for each random tree, and evaluate and analyze the decision tree. For each node in the tree, randomly select k feature components based on this point, and target different categories of features. Samples, assign different weights to find the best segmentation method;

(e) Divide the node into two branches according to the feature node with the best classification effect, and then recursively call step (d) until the tree can accurately classify the training sample set, or all attributes have been used;

(f) Repeat steps (c)-(e) until all m decision trees are established to generate a random forest model.

The prediction process of the random forest model:

(g) inputting the obtained user data (including gender, age, consumption level, income level, investment experience, registration days, click operation, browsing operation, consumption record) as independent variables into the random forest model;

(h) The majority voting method is used to comprehensively determine the classification results of multiple decision trees, and the risk classification results based on the random forest model are obtained. Specifically, the risk classification result includes each classification result and a classification result probability corresponding to each classification result.

In a specific embodiment, the risk includes: the probability that the user is the first classification result (conservative) is a, the probability that the user is the second classification result (robust) is b, and the user is The probability of the third classification result (balanced type) is c, the probability of the user being the fourth classification result (growth type) is d, and the probability of the user being the fifth classification result (aggressive type) is e.

The liquidity positioning module 203 is configured to perform liquidity positioning on the user according to the user profile, and obtain a liquidity positioning result of the user.

Specifically, the fluidity positioning result includes a bullet type, a step type, a term type, and a dumbbell type. Among them, the dumbbell type is used to indicate the types of users who mainly invest in short-term financial products and long-term financial products to weaken the investment of medium-term debt; the bullet type is used to indicate that the repayment period is highly concentrated on the yield curve User type at a certain point; the ladder type is used to indicate the type of users whose long-term, medium-term, and short-term wealth management products have basically the same proportion; the term type is used to indicate that they only prefer long-term, medium-term, and short-term financial products One of the user types.

In an embodiment, the step of positioning the user for mobility further includes:

According to the user profile of the user, a historical purchase record of the user is obtained. The user profile may include a user name, a registered email address, a registered mobile phone number, and an ID card number of the user. The historical purchase record is a purchase order for a financial product, and the purchase order can be used to find related orders in various financial systems through the user's username, registered mailbox, registered mobile phone number, and ID number. In an embodiment, when the number of purchase orders of the user is less than one, the user is prompted to conduct self-evaluation of his own liquidity, and obtain the user's liquidity self-evaluation information through an interactive interface and use it as a liquidity positioning result. .

The historical purchase records are analyzed to obtain product term information for each order. The product term information includes expiration date, value date and term length (short-term products, medium-term products, long-term products).

Determine the liquidity positioning result of the user according to the product term information of each order. In an embodiment, when the expiration date of each order is concentrated within 30 days and the value date is not concentrated within 30 days, it is determined that the user's liquidity positioning result is a bullet type; when the user's multiple When long-term and short-term products are included in each order but not medium-term products, the user's liquidity positioning result is determined to be dumbbell-shaped; when multiple orders of the user only include short-term products, the user's liquidity is determined The positioning result is term type (prefer short term); when multiple orders of the user only include mid-term products, it is determined that the liquidity positioning result of the user is term type (prefer middle term); when multiple orders of the user When only long-term products are included, it is determined that the user's liquidity positioning result is a term type (preferred long-term); when the above-mentioned judgment logic is not satisfied, it is determined that the user's liquidity positioning result is a step type.

The recommendation module 204 is configured to obtain the minimum investment amount of the user, and combine the risk classification result and the liquidity positioning result to recommend investment products suitable for the user.

The minimum investment amount can be obtained through user data, that is, calculated by obtaining a user's income level, consumption level, and consumption record. In another embodiment, it may be obtained through a user's interactive input on the APP.

In one embodiment, the user is provided with financial advice for six consecutive months based on the risk classification result and the liquidity positioning result.

In a specific embodiment, if there is a user B, male gender, age 50, income level annual salary of 100,000 to 240,000, average monthly consumption level of 50,000 to 10,000, preferred investment period is 0-3 months, and investment experience is 1 -3 years, mainly investing in financial management with capital protection and income. On average, it will be active once a day on the One Account App with 180 days of holding positions. It prefers current products and does not prefer regular products. It is predicted by the random forest model that User B ’s Risk tolerance is balanced and suitable for purchasing financial products with medium risk. According to user B's historical transaction records, it can be seen that most of User B's funds are used to invest in short-term wealth management products, and a small amount of funds are used to invest in regular wealth management products. It is inferred that User B is a dumbbell user. According to the configuration of the wealth management product currently held by user B, and combined with the current investment amount of user B of 20,000 yuan, it is recommended that user B invest in short-term wealth management products with a medium degree of risk.

Referring to FIG. 3, it is a program module diagram of the second embodiment of the financial product recommendation system 200 of the present application. In this embodiment, the financial product recommendation system 200 includes a conversion module 205 in addition to the collection module 201, the risk classification module 202, the liquidity positioning module 203, and the recommendation module 204 in the first embodiment.

The conversion module 205 is configured to convert the risk classification result into a continuity score. The risk classification result includes each classification result and a classification result probability corresponding to each classification result.

Specifically, the step of converting the risk classification result into a continuity score specifically includes:

(1) Obtain the classification result probability of each classification result and each classification result of the forest model. The risk classification results may include conservative type (probability a), robust type (probability b), and balanced type (probability is c) Growth type (probability d) and aggressive type (probability e).

(2) setting a mapping relationship between a risk classification result of the forest model and a mapping score;

(3) Combining the classification result probability with the mapping score to obtain the continuity score.

Specifically, the continuity score P is calculated as follows:

Among them, the classification result with the highest probability in each classification result is taken as the current classification result, ave_left_pro is the average probability of each level on the left side of the current classification result, ave_right_pro is the average probability of each level on the right side of the current classification result, and cur_pro is the probability of the current classification result .

In an embodiment, the output of a random forest of a user is level 3, and the mapped score is p = 5. The model predicts the probability of each level as a = 0.1, b = 0.1, c = 0.4, d = 0.2, and e = 0.2. The average left probability of the prediction level is ave_left_pro = (a + b) /2=0.1, and the average right probability is ave_right_pro = (d + e) /2=0.2. The continuity score of this user is:

In a further embodiment, when the risk classification result of the user is level 5 and the classification result probability e at the level is e.g. 0.5, the continuity score P of the user is obtained according to the following table:

Classification result probability e	(0.5,0.6)	(0.6,0.7)	(0.7,0.8)	[0.8,1]
Continuity score P	9.2	9.5	9.8	10

In addition, this application also proposes a method for recommending wealth management products.

Refer to FIG. 4, which is a schematic flowchart of a first embodiment of a financial product recommendation method of the present application. In this embodiment, according to different requirements, the execution order of the steps in the flowchart shown in FIG. 5 may be changed, and some steps may be omitted.

In step S400, user data of a user is collected, and the user data includes behavior data and user attributes.

In one embodiment, the collection module 201 may obtain the user profile from application software. The application software may be an account communication APP.

Specifically, the user attributes include gender, age, consumption level, income level, investment experience, and registration days.

The behavior data includes a user's operation record on the mobile terminal, and specifically includes a user's operation record on the One Account App, and the operation record may include a click operation, a browsing operation, and a consumption record.

Step S402: Establish a sample set according to the effective user data, and classify the risk capability of the user based on a random forest model to obtain a risk classification result of the user.

Specifically, the risk classification result may include a conservative type, a stable type, a balanced type, a growth type, and an aggressive type.

The effective user information is collected user information of multiple groups of users and corresponding risk classification results, wherein the user information includes gender, age, consumption level, income level, investment experience, registration days, click operation, and browse Features such as operations, consumption records are used as independent variables of the sample set, and the risk classification results are used as dependent variables of the sample set.

Further, the specific steps of classifying the risk capability of the user to obtain the risk classification result of the user will be described in detail in the second embodiment (FIG. 5) of the financial product product recommendation method of the present application.

Step S404: Locate the user's liquidity according to the user profile, and obtain the user's liquidity positioning result.

Specifically, the fluidity positioning result includes a bullet type, a step type, a term type, and a dumbbell type. Among them, the dumbbell type is used to indicate the types of users who mainly invest in short-term financial products and long-term financial products to weaken the investment of medium-term debt; the bullet type is used to indicate that the repayment period is highly concentrated on the yield curve User type at a certain point; the ladder type is used to indicate the type of users whose long-term, medium-term, and short-term wealth management products have basically the same proportion; the term type is used to indicate that they only prefer long-, medium-, and short-term financial products One of the user types.

Further, the specific steps of positioning the user's liquidity will be described in detail in the third embodiment (FIG. 6) of the financial product product recommendation method of the present application.

Step S406: Obtain the user's minimum investment amount, and combine the risk classification result and the liquidity positioning result to recommend investment products suitable for the user.

The minimum investment amount can be obtained through user data, that is, calculated by obtaining a user's income level, consumption level, and consumption record. In another embodiment, it may be obtained through a user's interactive input on the APP.

In one embodiment, the user is provided with financial advice for six consecutive months based on the risk classification result and the liquidity positioning result.

In a specific embodiment, if there is a user B, male gender, age 50, income level annual salary of 100,000 to 240,000, average monthly consumption level of 50,000 to 10,000, preferred investment period is 0-3 months, and investment experience is 1 -3 years, mainly investing in financial management with capital protection and income. On average, it will be active once a day on the One Account App with 180 days of holding positions. It prefers current products and not regular products. It is predicted by the random forest model. Risk tolerance is balanced and suitable for purchasing financial products with medium risk. According to user B's historical transaction records, it can be seen that most of User B's funds are used to invest in short-term wealth management products, and a small amount of funds are used to invest in regular wealth management products. According to the configuration of the wealth management product currently held by user B, and combined with the current investment amount of user B of 20,000 yuan, it is recommended that user B invest in short-term wealth management products with a medium degree of risk.

As shown in FIG. 5, it is a schematic flowchart of a second embodiment of a method for recommending wealth management products in this application. In this embodiment, according to different requirements, the execution order of the steps in the flowchart shown in FIG. 5 may be changed, and some steps may be omitted.

In this embodiment, the step of classifying the risk capability of the user to obtain a risk classification result of the user specifically includes:

Random forest model training process: S500, establish a sample set according to valid user data;

S502. Generate n feature samples by resampling from the sample set. The number of feature components in each feature sample is a, the number of decision trees is m, and the number of decision features of each decision tree is k;

S504. Use the Bagging algorithm to sample n feature samples m times to obtain m feature sets.

S506. A feature set is randomly selected for each random tree, and the decision tree is evaluated and error analyzed. For each node in the tree, k feature components based on this point are randomly selected, and feature samples of different categories are targeted. To assign different weights to find the best segmentation method;

S508. Divide the node into two branches according to the feature node with the best classification effect, and then call S506 recursively until the tree can accurately classify the training sample set, or all attributes have been used;

S510. Repeat S504-S508 until all m decision trees are established to generate a random forest model.

The prediction process of the random forest model:

S512. Input the obtained user data (including gender, age, consumption level, income level, investment experience, registration days, click operation, browsing operation, consumption record) as independent variables into the random forest model.

S514. The majority voting method is used to comprehensively determine the classification results of multiple decision trees, and the risk classification results based on the random forest model are obtained. Specifically, the risk classification result includes each classification result and a classification result probability corresponding to each classification result.

In a specific embodiment, the risk includes: the probability that the user is the first classification result (conservative) is a, the probability that the user is the second classification result (robust) is b, and the user is The probability of the third classification result (balanced type) is c, the probability of the user being the fourth classification result (growth type) is d, and the probability of the user being the fifth classification result (aggressive type) is e.

As shown in FIG. 6, it is a schematic flowchart of a third embodiment of a method for recommending wealth management products in this application. . In this embodiment, according to different requirements, the execution order of the steps in the flowchart shown in FIG. 6 may be changed, and some steps may be omitted.

In this embodiment, the step of positioning the user for mobility specifically includes:

S600. Acquire a historical purchase record of the user according to the user profile of the user.

The user profile may include a user name, a registered email address, a registered mobile phone number, and an ID card number of the user. The historical purchase record is a purchase order for a financial product, and the purchase order can be used to find related orders in various financial systems through the user's username, registered mailbox, registered mobile phone number, and ID number. In an embodiment, when the number of purchase orders of the user is less than one, the user is prompted to conduct self-evaluation of his own liquidity, and obtain the user's liquidity self-evaluation information through an interactive interface and use it as a liquidity positioning result .

S602. Analyze the historical purchase records to obtain product term information of each order.

The product term information includes expiration date, value date and term length (short-term products, medium-term products, long-term products).

S604. Determine the liquidity positioning result of the user according to the product term information of each order.

In an embodiment, when the expiration date of each order is concentrated within 30 days and the value date is not concentrated within 30 days, it is determined that the user's liquidity positioning result is a bullet type; when the user's multiple When long-term and short-term products are included in each order but not medium-term products, the user's liquidity positioning result is determined to be dumbbell-shaped; when multiple orders of the user only include short-term products, the user's liquidity is determined The positioning result is term type (prefer short term); when multiple orders of the user only include mid-term products, it is determined that the liquidity positioning result of the user is term type (prefer middle term); when multiple orders of the user When only long-term products are included, it is determined that the user's liquidity positioning result is a term type (preferred long-term); when the above-mentioned judgment logic is not satisfied, it is determined that the user's liquidity positioning result is a step type.

As shown in FIG. 7, it is a schematic flowchart of a fourth embodiment of a method for recommending wealth management products of the present application. In this embodiment, steps S700-S702 of the financial product recommendation method are similar to steps S400-S402 of the first embodiment, except that the method further includes steps S704-S708.

The risk classification result includes each classification result and a classification result probability corresponding to each classification result.

Step S704: Obtain each classification result of the forest model and the classification result probability corresponding to each classification result. The risk classification result may include conservative type (probability a), robust type (probability b), and balanced type (probability is c) growth type (probability d) and aggressive type (probability e);

Step S706: Set a mapping relationship between a risk classification result of the forest model and a mapping score;

Step S708: Combine the classification result probability on the basis of the mapping score to obtain the continuity score.

Specifically, the continuity score P is calculated as follows:

Among them, the classification result with the highest probability in each classification result is taken as the current classification result, ave_left_pro is the average probability of each level on the left side of the current classification result, ave_right_pro is the average probability of each level on the right side of the current classification result, and cur_pro is the probability of the current classification result .

In an embodiment, the output of a random forest of a user is level 3, and the mapped score is p = 5. The model predicts the probability of each level as a = 0.1, b = 0.1, c = 0.4, d = 0.2, and e = 0.2. The average left probability of the prediction level is ave_left_pro = (a + b) /2=0.1, and the average right probability is ave_right_pro = (d + e) /2=0.2. The continuity score of this user is:

In a further embodiment, when the risk classification result of the user is level 5 and the classification result probability e at the level is e.g. 0.5, the continuity score P of the user is obtained according to the following table:

Classification result probability e	(0.5,0.6)	(0.6,0.7)	(0.7,0.8)	[0.8,1]
Continuity score P	9.2	9.5	9.8	10

The above-mentioned serial numbers of the embodiments of the present application are merely for description, and do not represent the superiority or inferiority of the embodiments.

Through the description of the above embodiments, those skilled in the art can clearly understand that the methods in the above embodiments can be implemented by means of software plus a necessary universal hardware platform, and of course, also by hardware, but in many cases the former is better. Implementation. Based on such an understanding, the technical solution of this application that is essentially or contributes to the existing technology can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, The optical disc) includes several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in the embodiments of the present application.

The above are only preferred embodiments of the present application, and thus do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the contents of the specification and drawings of the present application, or directly or indirectly used in other related technical fields Are included in the scope of patent protection of this application.

Claims (20)

1. A method for recommending financial products, which is applied to a server, is characterized in that the method includes steps:

   Collect user data of users;

   Establish a sample set according to valid user data, classify the user's risk ability based on a random forest model, and obtain a risk classification result for the user;

   Locating the user's mobility according to the user profile to obtain a result of the user's mobility positioning; and

   Obtain the user's minimum investment amount, and combine the risk classification result and the liquidity positioning result to recommend investment products suitable for the user.
2. The method for recommending financial products according to claim 1, wherein the step of establishing a sample set based on valid user data, classifying the user's risk ability based on a random forest model, and obtaining the risk classification result of the user is specific include:

   (a) establishing a sample set based on valid user data;

   (b) Generate n feature samples from the sample set by resampling. The number of feature components in each feature sample is a, the number of decision trees is m, and the number of decision features of each decision tree. K

   (c) Sample the n feature samples m times using the Bagging algorithm to obtain m feature sets;

   (d) Randomly select a feature set for each random tree, and evaluate and analyze the decision tree. For each node in the tree, randomly select k feature components based on this point, and target different categories of features. Samples, assign different weights to find the best segmentation method;

   (e) Divide the node into two branches according to the feature node with the best classification effect, and then recursively call step (d) until the tree can accurately classify the training sample set, or all attributes have been used;

   (f) repeating steps (c)-(e) until all m decision trees are established to generate a random forest;

   (g) inputting the obtained user data as an independent variable into the random forest model;

   (h) The majority voting method is used to comprehensively determine the classification results of multiple decision trees, and the risk classification results based on the random forest model are obtained.
3. The method for recommending financial products according to claim 1, further comprising the steps of: after classifying the risk ability of the user based on a random forest model to obtain the risk classification result of the user:

   The risk classification result is converted into a continuity score.
4. The method for recommending financial products according to claim 2, characterized in that after classifying the risk capability of the user based on a random forest model and obtaining the risk classification result of the user, the method further comprises the steps of:

   The risk classification result is converted into a continuity score.
5. The financial product recommendation method according to claim 3, wherein the risk classification result comprises each classification result and a classification result probability corresponding to each classification result, and the risk classification result is converted into a continuity score The steps include:

   Obtaining each classification result and the classification result probability corresponding to each classification result of the forest model;

   Setting a mapping relationship between a risk classification result of the forest model and a mapping score;

   The continuity score is obtained by combining the classification result probability on the basis of the mapping score.
6. The financial product recommendation method according to claim 4, wherein the risk classification result includes each classification result and a classification result probability corresponding to each classification result, and the risk classification result is converted into a continuity score The steps include:

   Obtaining each classification result and the classification result probability corresponding to each classification result of the forest model;

   Setting a mapping relationship between a risk classification result of the forest model and a mapping score;

   The continuity score is obtained by combining the classification result probability on the basis of the mapping score.
7. The financial product recommendation method according to claim 1, wherein the step of performing liquidity positioning on the user according to the user profile to obtain the liquidity positioning result of the user specifically includes:

   Obtaining historical purchase records of the user according to the user profile;

   Analyze the historical purchase records to obtain product term information for each order;

   Determine the liquidity positioning result of the user according to the product term information of each order.
8. A server, characterized in that the server includes a memory and a processor, and the memory stores a financial product recommendation system operable on the processor, and when the financial product recommendation system is executed by the processor To achieve the following steps:

   Collect user data of users;

   Establish a sample set according to valid user data, classify the user's risk ability based on a random forest model, and obtain a risk classification result for the user;

   Locating the user's mobility according to the user profile to obtain a result of the user's mobility positioning; and

   Obtain the user's minimum investment amount, and combine the risk classification result and the liquidity positioning result to recommend investment products suitable for the user.
9. The server according to claim 8, wherein the step of establishing a sample set according to valid user data, classifying the risk capability of the user based on a random forest model, and obtaining the risk classification result of the user specifically comprises:

   (a) Establish a sample set based on valid user data, and a sample set based on the data;

   (b) Generate n feature samples from the sample set by resampling. The number of feature components in each feature sample is a, the number of decision trees is m, and the number of decision features of each decision tree. K

   (c) Sample the n feature samples m times using the Bagging algorithm to obtain m feature sets;

   (d) Randomly select a feature set for each random tree, and evaluate and analyze the decision tree. For each node in the tree, randomly select k feature components based on this point, and target different categories of features Samples, assign different weights to find the best segmentation method;

   (e) Divide the node into two branches according to the feature node with the best classification effect, and then recursively call step (d) until the tree can accurately classify the training sample set, or all attributes have been used;

   (f) repeating steps (c)-(e) until all m decision trees are established to generate a random forest;

   (g) inputting the obtained user data as an independent variable into the random forest model;

   (h) The majority voting method is used to comprehensively determine the classification results of multiple decision trees, and the risk classification results based on the random forest model are obtained.
10. The server according to claim 8, characterized in that after classifying the risk capability of the user based on a random forest model and obtaining the risk classification result of the user, further comprising the step of:

    The risk classification result is converted into a continuity score.
11. The server according to claim 9, characterized in that after classifying the risk capability of the user based on a random forest model and obtaining the risk classification result of the user, the method further comprises the step of:

    The risk classification result is converted into a continuity score.
12. The server according to claim 10, wherein the risk classification result includes each classification result and a classification result probability corresponding to each classification result, and the step of converting the risk classification result into a continuity score is specific include:

    Obtaining each classification result and the classification result probability corresponding to each classification result of the forest model;

    Setting a mapping relationship between a risk classification result of the forest model and a mapping score;

    The continuity score is obtained by combining the classification result probability on the basis of the mapping score.
13. The server according to claim 11, wherein the risk classification result includes each classification result and a classification result probability corresponding to each classification result, and the step of converting the risk classification result into a continuity score is specific include:

    Obtaining each classification result and the classification result probability corresponding to each classification result of the forest model;

    Setting a mapping relationship between a risk classification result of the forest model and a mapping score;

    The continuity score is obtained by combining the classification result probability on the basis of the mapping score.
14. The server according to claim 8, wherein the step of performing liquidity positioning on the user according to the user profile to obtain the liquidity positioning result of the user specifically includes:

    Obtaining historical purchase records of the user according to the user profile;

    Analyze the historical purchase records to obtain product term information for each order;

    Determine the liquidity positioning result of the user according to the product term information of each order.
15. A computer-readable storage medium stores a financial product recommendation system, and the financial product recommendation system is executable by at least one processor, so that the at least one processor performs the following steps:

    Collect user data of users;

    Establish a sample set according to valid user data, classify the user's risk ability based on a random forest model, and obtain a risk classification result for the user;

    Locating the user's mobility according to the user profile to obtain a result of the user's mobility positioning; and

    Obtain the user's minimum investment amount, and combine the risk classification result and the liquidity positioning result to recommend investment products suitable for the user.
16. The computer-readable storage medium of claim 15, wherein the step of establishing a sample set based on valid user data, classifying the user's risk ability based on a random forest model, and obtaining a risk classification result for the user These include:

    (a) Establish a sample set based on valid user data, and a sample set based on the data;

    (b) Generate n feature samples from the sample set by resampling. The number of feature components in each feature sample is a, the number of decision trees is m, and the number of decision features of each decision tree. K

    (c) Sample the n feature samples m times using the Bagging algorithm to obtain m feature sets;

    (d) Randomly select a feature set for each random tree, and evaluate and analyze the decision tree. For each node in the tree, randomly select k feature components based on this point, and target different categories of features. Samples, assign different weights to find the best segmentation method;

    (e) Divide the node into two branches according to the feature node with the best classification effect, and then recursively call step (d) until the tree can accurately classify the training sample set, or all attributes have been used;

    (f) repeating steps (c)-(e) until all m decision trees are established to generate a random forest;

    (g) inputting the obtained user data as an independent variable into the random forest model;

    (h) The majority voting method is used to comprehensively determine the classification results of multiple decision trees, and the risk classification results based on the random forest model are obtained.
17. The computer-readable storage medium according to claim 15, wherein after classifying the risk capability of the user based on a random forest model and obtaining the risk classification result of the user, further comprising the step of:

    The risk classification result is converted into a continuity score.
18. The computer-readable storage medium of claim 16, further comprising: after the risk capability classification of the user is performed based on a random forest model to obtain the risk classification result of the user, further comprising the step of:

    The risk classification result is converted into a continuity score.
19. The computer-readable storage medium according to claim 17 or 18, wherein the risk classification result comprises each classification result and a classification result probability corresponding to each classification result, and the risk classification result is converted into The steps of the continuity score include:

    Obtaining each classification result and the classification result probability corresponding to each classification result of the forest model;

    Setting a mapping relationship between a risk classification result of the forest model and a mapping score;

    The continuity score is obtained by combining the classification result probability on the basis of the mapping score.
20. The computer-readable storage medium according to claim 15, wherein the step of performing liquidity positioning on the user according to the user profile to obtain the liquidity positioning result of the user specifically includes:

    Obtaining historical purchase records of the user according to the user profile;

    Analyze the historical purchase records to obtain product term information for each order;

    Determine the liquidity positioning result of the user according to the product term information of each order.

Images