WO2019061976A1

WO2019061976A1 - Fund product recommendation method and apparatus, terminal device, and storage medium

Info

Publication number: WO2019061976A1
Application number: PCT/CN2018/074570
Authority: WO
Inventors: 姬马婧雯; 何军
Original assignee: 平安科技（深圳）有限公司
Priority date: 2017-09-28
Filing date: 2018-01-30
Publication date: 2019-04-04
Also published as: CN107679946A; CN107679946B

Abstract

The present application discloses a fund product recommendation method and apparatus, a terminal device, and a storage medium. Said fund product recommendation method comprises: acquiring current user profile data, the current user profile data comprising at least one piece of current characteristic data; acquiring a user data model, the user data model comprising at least two clustered clusters, each clustered cluster corresponding to a risk evaluation value; on the basis of the current user profile data, acquiring from the at least two clustered clusters a target clustered cluster corresponding to the at least one piece of current characteristic data; and on the basis of a risk evaluation value corresponding to the target clustered cluster, determining a target fund product. Said fund product recommendation method can recommend a fund product on the basis of the investment condition of a user, improving the accuracy of positioning of fund products made by investors.

Description

Fund product recommendation method, device, terminal device and storage medium

This patent application is based on the Chinese invention patent application filed on September 28, 2017, with the application number of 201710899347.8, entitled "Recommended method, device, terminal equipment and storage medium based on fund products", and requires priority.

Technical field

The present application relates to the field of financial data processing, and in particular, to a fund product recommendation method, device, terminal device and storage medium.

Background technique

Fund investment refers to a kind of financial management means for investors to purchase fund products through the fund trading system to realize the management and distribution of assets. In the current fund trading system, according to the level of investment risk, the fund products are divided into fund types such as stock type, index type, hybrid type, bond type and currency type. Users select fund products of different fund types according to fund type and investment conditions. transaction. The current fund trading system does not have the function of recommending fund products based on the user's investment conditions, which makes investors' investment conditions and fund products to be positioned, which affects the return rate of investment funds due to low positioning accuracy.

Summary of the invention

The application provides a fund product recommendation method, device, terminal device and storage medium to solve the problem that the current fund transaction system does not have the fund product recommendation based on the user's investment condition.

In a first aspect, the present application provides a fund product recommendation method, including:

Obtaining current user portrait data, the current user portrait data including at least one current feature data;

Obtaining a user data model, where the user data model includes at least two clustering clusters, each of the clustering clusters corresponding to a risk assessment value;

And acquiring, according to the current user portrait data, a target clustering cluster corresponding to at least one of the current feature data from at least two of the clustering clusters;

The target fund product is determined based on the risk assessment value corresponding to the target clustering cluster.

In a second aspect, the application provides a fund product recommendation device, including:

a current user image data obtaining module, configured to acquire current user portrait data, where the current user portrait data includes at least one current feature data;

a user data model obtaining module, configured to acquire a user data model, where the user data model includes at least two clustering clusters, each of the clustering clusters corresponding to a risk assessment value;

a target clustering cluster determining module, configured to acquire, according to the current user portrait data, a target clustering cluster corresponding to at least one of the current feature data from at least two of the clustering clusters;

The target fund product determining module is configured to determine the target fund product based on the risk assessment value corresponding to the target clustering cluster.

In a third aspect, the present application provides a terminal device including a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, the processor executing the computer The following steps are implemented when reading the instruction:

In a fourth aspect, the present application provides a computer readable storage medium storing computer readable instructions that, when executed by a processor, implement the following steps:

In the fund product recommendation method, device, device and storage medium provided by the present application, the target risk assessment value is determined based on the current user portrait data and the user data type, and the target target fund product is determined by using the target risk assessment value, so that The target fund products recommended to the target users accurately locate the user's own investment conditions and fund products, which helps the target users to improve the investment return rate.

DRAWINGS

In order to more clearly illustrate the technical solutions of the present application, the drawings to be used in the description of the present application will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the present application. For ordinary technicians, other drawings can be obtained based on these drawings without paying for creative labor.

1 is a flow chart of a fund product recommendation method in Embodiment 1.

2 is another flow chart of the fund product recommendation method in Embodiment 1.

FIG. 3 is a specific flow chart of step S50 of FIG. 2.

4 is a specific flow chart of step S30 of FIG. 1.

FIG. 5 is a specific flowchart of step S40 in FIG. 1.

6 is a schematic block diagram of a fund product recommendation device in Embodiment 2.

Figure 7 is a schematic diagram of a terminal device in Embodiment 4.

Detailed ways

The technical solutions in the present application will be clearly and completely described in the following with reference to the drawings in the present application. It is obvious that the described embodiments are a part of the embodiments of the present application, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present application without departing from the inventive scope are the scope of the present application.

Example 1

FIG. 1 shows a flow chart of a fund product recommendation method in this embodiment. The fund product recommendation method is applied in the fund trading system, and is used to recommend the target fund product to the user according to the investment condition of the user, so as to avoid the investor's low investment accuracy of the investment condition and the fund product, which affects the profit rate of the investment fund. As shown in Figure 1, the fund product recommendation method includes the following steps:

S10: Acquire current user portrait data, and the current user portrait data includes at least one current feature data.

Among them, the current user portrait data is user portrait data of a user who needs to recommend a fund product. In this embodiment, the user who needs to recommend the fund product is simply referred to as the target user. User portraits are a tagged user model abstracted from information such as user social attributes, lifestyle habits, and consumer behavior. The user portrait data is data for constructing the user portrait. The current user portrait data can reflect the investment conditions of the target user.

The current feature data is data related to the target user's own investment conditions. The current feature data includes, but is not limited to, age, occupation, income, investment experience, investment ratio, risk preference, and bearish loss value in the present embodiment. That is, the current user portrait data includes at least one current feature data, specifically, at least one of age, occupation, income, investment experience, investment ratio, risk preference, and loss tolerance value.

The age in the current feature data is the age of the target user, and the age of the target user is related to the fund type of the fund product that the target user may purchase. Generally speaking, the older the target user is, the more biased it is to purchase a fund product with lower risk. Conversely, the younger the target user is, the more biased it is to purchase a fund product with higher risk. The occupation in the current feature data is the occupation of the target user. The occupation of the target user is generally related to the personality of the target user. The target users with different personalities may select fund products with different risks; otherwise, the target users with the same personality may choose to purchase the same risk. Fund products. The income in the current characteristic data may be the disposable income of the target user, or the net income of the target user, and the income may also affect the fund type of the target user selecting the fund product. The investment experience in the current characteristic data refers to the experience of the target user in investing in funds or investing in other financial wealth management products. The less general investment experience, the more biased it is to purchase fund products with lower risk. The proportion of investment in the current characteristic data refers to the ratio of the target user's purchase of the fund product to all the investment. According to the proportion of the investment, it can be determined which kind of fund product the user prefers to purchase. The risk preference in the current feature data is the risk preference of the target user, and the fund products of different risks are recommended according to the risk preference. The value of the loss in the current characteristic data is the value that the target user can bear the loss, and the fund product with different risk can be recommended based on the loss value.

S20: Acquire a user data model, where the user data model includes at least two clustering clusters, and each clustering cluster corresponds to a risk assessment value.

The user data model is a model in which the fund trading system obtains a risk assessment value associated with an investment fund after training based on training user portrait data. The training user portrait data is user image data of the training user, and the training user portrait data is data for training the user data model. The training user is a user who has previously opened an account in the fund trading system and conducted a fund transaction. The clustering cluster is a collection of similar training user image data obtained by clustering the training user image data by using a clustering algorithm. Each cluster cluster corresponds to a risk assessment value, which is an evaluation value of the risk that the training user corresponding to the training user portrait data in any cluster cluster can bear the risk when investing in the fund. Understandably, the risk assessment value is related to the fund type of the fund product. The training user with the higher risk assessment value tends to invest in the fund product with high risk and high return. On the contrary, the less the risk assessment value, the more the training user is biased. Fund products with low investment risk and low returns.

In a specific embodiment, as shown in FIG. 2, the fund product recommendation method further includes:

S50: Train the user data model based on the training user image data, and the training user image data includes at least one training feature data.

Among them, the training feature data is data related to training the user's own investment conditions. The training characteristic data includes, but is not limited to, age, occupation, income, investment experience, investment ratio, risk preference, and loss tolerance value in the present embodiment. That is, the current user portrait data includes at least one training feature data, specifically, at least one of age, occupation, income, investment experience, investment ratio, risk preference, and loss tolerance value. Specifically, training the user data model based on the training user portrait data refers to clustering at least one training feature data in the training user portrait data, and using the set of similar training user portrait data as a cluster cluster to image all the training users. The data is divided into at least two cluster clusters, and the risk assessment values corresponding to each cluster cluster are obtained, thereby forming a user data model.

The age in the training feature data is the age of the training user, and the age of the training user is related to the fund type of the fund product that the training user may purchase. Generally speaking, the older the training user is, the more he prefers to buy a fund product with lower risk. On the contrary, the younger the training user, the more he prefers to buy a higher risk fund product. The occupation in the training characteristic data is the occupation of the training user. The occupation of the training user is generally related to the personality of the training user. The training users with different personality may select the fund products with different risks; on the contrary, the training users with the same personality may choose to purchase the same risk. Fund products. The income in the training characteristic data may be the disposable income of the training user, or the net income of the training user, and the income may also affect the fund type of the training user to select the fund product. The investment experience in the training characteristic data refers to the experience of training users in investing in funds or investing in other financial wealth management products. The less general investment experience, the more biased it is to purchase fund products with lower risk. The proportion of investment in the training characteristic data refers to the ratio of training users to purchase fund products in all investments. According to the proportion of investment, it can determine which kind of risk fund products the user prefers to purchase. The risk preference in the training feature data is to train the user's risk preference, and recommend different risk fund products according to the risk preference. The value of the loss in the training characteristic data is the value that the training user can bear the loss, and the fund product with different risk can be recommended based on the loss value.

In this embodiment, as shown in FIG. 3, in step S50, the user data model is trained based on the training user image data, and specifically includes the following steps:

S51: Perform normalization processing on at least one training feature data in the training user portrait data, so that the training user image data includes at least one standard feature data.

Among them, data normalization is to scale the data to a small specific interval, to remove the unit limit of the data, and convert it into a pure value of the infinite level, which is convenient for different units or magnitude indicators. Ability to compare and weight. Specifically, when normalizing at least one training feature data in the training user portrait data, a conversion process may be performed using min-max normalization to acquire at least one standard feature data. Among them, min-max normalization is also called deviation normalization, which refers to the process of linearly transforming the original data by using a conversion function to make the result fall into a preset interval, wherein the conversion function

Min is the minimum value of the sample data, max is the maximum value of the sample data, and N is the interval size of the preset interval. It can be understood that if N is 1, the result after the min-max normalization process falls within the range of [0, 1]; if N is 10, the result after the min-max normalization process falls in [0] , 10] within this range.

S52: K-means clustering algorithm is used to cluster at least one standard feature data in the training user image data, and at least two clustering clusters are obtained, and each clustering cluster corresponds to a centroid user image data.

Among them, K-means clustering algorithm is a clustering algorithm based on distance evaluation similarity, that is, the closer the distance between two objects, the larger the similarity is. The K-means clustering algorithm evaluates the similarity of two objects according to the Euclidean distance by calculating the Euclidean distance of two objects. Euclidean metric (also known as Euclidean metric) refers to the true distance between two points in m-dimensional space, or the natural length of the vector (ie, the distance from the point to the origin). Euclidean distance of any two n-dimensional vectors a(X _i1 ,X _i2 ,...,X _in ) and b(X _j1 ,X _j2 ,...,X _jn )

The training user portrait data is user portrait data of a training user for training a user data model, and the training user portrait data includes at least one training feature data. Step S51: Convert each training feature data in the training user image data into standard feature data, so that in step S52, the K-means clustering algorithm is used to cluster at least one standard feature data in the training user image data to obtain at least Two clustering clusters, each of which includes training user portrait data corresponding to a plurality of training users. In any of the clustering clusters, there is a centroid user image data corresponding to the centroid user in the training user image data of the plurality of training users, so that the sum of the distances of the other training user image data to the centroid user portrait data is minimized. It can be understood that the centroid user portrait data is one of all training user portrait data in any cluster cluster, and therefore, the centroid user portrait data also includes at least one standard feature data converted from the training feature data. Step S52 will be described in detail below with reference to specific examples.

First, the training user image data of the m training users is acquired, and at least one training feature data in the training user image data is normalized to form at least one standard feature data, and the user data matrix R is obtained (as shown in the following table). The user data matrix R is an m*n matrix, m is the number of training users, and n is the number of standard feature data. When the training characteristic data of the age is normalized, if the minimum value of the age is 18, the maximum value is 88, and the size of N is 10, if the age of a training user is 48, the standard characteristic data acquired by the conversion function is used. Is 5. When standardizing the training characteristic data of the occupation, the comparison table of the occupational and standard characteristic data may be preset, and the comparison table may be obtained based on the training characteristic data of the occupation in the training user portrait data, and the corresponding standard feature may be obtained. data. Wherein, in the comparison table, different standardized scores may be enumerated according to occupations, and each standardized score corresponds to different occupations of different enterprises, so as to query corresponding standardized scores based on the training characteristic data of the occupation in the training user portrait data, as Corresponding standard feature data. Similarly, other feature data is standardized by using a conversion function or a comparison table.

	U1U1	U2U2	U3U3	U4U4	U5U5	U6U6	……......	UmUm
年龄age	55	33	22	33	11	44	……......	77
职业Career	55	55	33	44	22	66	……......	66
收入income	66	44	11	66	44	77	……......	55
投资经验Investment experience	88	66	22	77	44	77	……......	66
投资比例Investment ratio	44	66	33	99	33	55	……......	44
风险偏好Risk preference	66	44	66	77	22	88	……......	77
承受亏损值Bear the loss	66	55	33	77	22	55	……......	88
……......	……......	……......	……......	……......	……......	……......	……......	……......

Then, the values of the standard feature data in the user data matrix R are clustered by a K-means clustering algorithm. The K-means clustering algorithm is used to perform the clustering process as follows: Step (1), an n-dimensional map is established, and m data points in the n-dimensional graph are drawn according to the values of the standard feature data of each training user Ui in the user data matrix R. Ui, where i∈m, each data point Ui corresponds to a training user. Step (2), pre-defining the K value, according to the K value, the m data points can be divided into K data sets G=[G1, G2, G3, G4, ... Gj..., Gk], where K≥2,j ∈k. In step (3), one data point Ui is randomly selected in each data set Gj as the centroid Ci such that there are K centroids Ci in all data sets. In step (4), the Euclidean distance Di of any of the data points Ui and the K centroids Gi in each data set Gj is calculated, and the data points Ui are classified into a data set Gj having the smallest Euclidean distance Di. In step (5), all data points Ui are executed in step (4) to form a new data set G. Repeat steps (3)-(5) such that when the new centroid Ci and the old centroid Ci in any data set Gj are less than a preset threshold, the K-means clustering algorithm terminates, forming K cluster clusters, each The cluster cluster has a centroid user, and the centroid user corresponds to the centroid user portrait data.

S53: weighting the centroid user image data by using a weighting algorithm to determine that the centroid user image data corresponds to a risk assessment value.

Wherein, the weighting algorithm is

among them,

Pi is the risk assessment value of the centroid user, Vi is the value of each standard feature data in the centroid user portrait data, and Wi is the weight of each standard feature data. The weight of each standard feature data may be a value obtained by statistically processing the standard feature data of each training user's portrait data in advance using a multivariate linear regression model, so as to be directly called when a risk assessment is required. The multivariate linear regression model is h _θ (x)=θ ₀ +θ ₁ x ₁ +θ ₂ x ₂ +...+θ _n x _n , where h _θ (x) is a hypothesis function and each θ is an input The angle vector between values, each x is a corresponding feature. When x ₀ is added to the above formula, x ₀ =1, then h _θ (x)=θ ₀ x ₀ +θ ₁ x ₁ +θ ₂ x ₂ + ... + θ _n x _n = θ ^T X. Where θ is the row vector, the row vector contains the parameters in the linear regression model, and X is the sample feature matrix.

S54: Acquire a user data model based on the clustering cluster and the risk assessment value.

In this embodiment, since the K-means clustering algorithm is used to divide all training user image data in the user data matrix R into K cluster clusters, the centroid user image data of each cluster cluster and the same cluster cluster The other training user image data is similar, and the risk evaluation value corresponding to the centroid user portrait data can be used as the risk evaluation value of other training user portrait data in the clustering cluster to determine the user data model.

S60: Store the user data model in a database.

In this embodiment, the user data model trained in step S50 is stored in MySQL, Oracle or other database, so that the pre-trained user data model is called from the database when the fund product needs to be recommended to the target user.

In this embodiment, step S20 includes: obtaining a user data model from a database. Since the user data model is pre-trained and stored in the database, when the user data model needs to be recommended for the fund product, the user data model can be directly called from the database, and the corresponding recommendation processing can be performed, and the operation process is simple and fast.

S30: Acquire, according to the current user portrait data, a target clustering cluster corresponding to the at least one current feature data from the at least two clustering clusters.

The target clustering cluster refers to the cluster cluster in which the centroid user portrait data closest to the current user image data is located. The target clustering cluster specifically refers to a clustering cluster corresponding to the centroid user portrait data closest to the current user portrait data formed by the at least one current feature data.

In a specific implementation, as shown in FIG. 4, step S30 specifically includes the following steps:

S31: Calculate the current user image data and the centroid user image data of at least two cluster clusters in the user data model to obtain at least two Euclidean distances.

In this embodiment, K cluster clusters are stored in the user data model, and each cluster cluster corresponds to a centroid user image data. If the current user image data is an n-dimensional vector a (X _i1 , X _i2 ,. .., X _in ), the centroid user image data of any cluster cluster is n-dimensional vector b (X _j1 , X _j2 ,..., X _jn ), then the current user image data and the centroid user image data are Distance

Wherein, the dimension n of the vector a corresponds to the number of current feature data in the current user portrait data; accordingly, the dimension n of the vector b corresponds to the number of training feature data in the centroid user portrait data.

It can be understood that, before step S31, at least one current feature data in the current user image data needs to be normalized to include at least one standard feature data in the current user image data, so as to calculate current user image data and a centroid user image. The Euclidean distance between the data helps to simplify the calculation process and improve the calculation efficiency. The process of normalizing at least one current training data in the current user image data is the same as the process of normalizing at least one training feature data in the training user image data. To avoid repetition, details are not described herein.

S32: Select a cluster cluster of the centroid user image data corresponding to the minimum value of the at least two Euclidean distances as the target cluster cluster corresponding to the at least one current feature data.

Since the Euclidean distance D _{a,b of the} K current user portrait data and the centroid user portrait data is obtained in step S31, the cluster class of the centroid user portrait data corresponding to the minimum value is selected from the K Euclidean distances D _a,b . The cluster is a target cluster class cluster corresponding to at least one current feature data in the current user portrait data.

S40: Determine the target fund product based on the risk assessment value corresponding to the target clustering cluster.

In this embodiment, the risk assessment value corresponding to the target cluster cluster is defined as the target risk assessment value. In step S52, the centroid user image data of each cluster cluster is weighted by a weighting algorithm to determine a risk assessment value corresponding to the centroid user portrait data, and the risk assessment value is used as a corresponding cluster cluster. Risk assessment value. In this embodiment, the target clustering cluster is one of at least two clustering clusters, and the risk assessment value corresponding to the target clustering cluster may be used as the target risk assessment value.

Among them, the target fund product is a fund product recommended by the fund trading system based on the obtained target risk assessment value to the target user. The target fund product is associated with the current user portrait data of the target user, so that the target fund product meets the investment condition of the target user; and the target fund product is recommended based on the target risk assessment value, so that it is related to the fund type of the fund product, The positioning of fund products is more accurate. In this embodiment, after obtaining the target fund product, the fund trading system displays the target fund product through a terminal such as a smart phone or a tablet computer, so that the target user understands the target fund product, and performs fund trading based on the target fund product to assist the user. Have a more accurate understanding of their own investment conditions and the positioning of fund products, which will help target users to improve their investment returns.

In a specific implementation, as shown in FIG. 5, step S40 specifically includes the following steps:

S41: Determine a corresponding target fund type based on the risk assessment value corresponding to the target clustering cluster.

Among them, the type of fund is based on the type of risk of the fund. In this embodiment, the fund types include, but are not limited to, stock type, index type, hybrid type, bond type, and currency type. In the fund trading system, the corresponding relationship between fund risk and fund type is preset. If the risk range of the fund is (0, 10), divide the risk risk interval (0, 2), (2, 4), (4, 6], (6, 8) and (8, 10) into five. The risk levels correspond to the five fund types of currency type, bond type, hybrid type, index type and stock type, and make the fund products of the fund trading system uniquely correspond to one type of fund. In this embodiment, it can be standardized. Therefore, the obtained target risk assessment value is within the interval (0, 10), so that the corresponding fund type is directly determined according to the target risk assessment value.

S42: According to the fund type, obtain the fund products and fund evaluation indicators corresponding to the fund type.

Among them, the fund products to be recommended are all fund products of the fund trading system whose fund type is the target fund type. The fund evaluation index is used to evaluate the quality of the fund product. The fund evaluation index is related to the investment income and/or investment risk of the fund product. Specifically, the fund evaluation index is associated with the fund type, and the fund evaluation index for evaluating the corresponding fund product may be determined based on the target fund type.

Specifically, the fund evaluation index includes, but is not limited to, an average return rate, an alpha coefficient, a standard deviation, a beta coefficient, a morning star risk coefficient, a Sharpe ratio, and an R square in the present embodiment. Among them, the average rate of return is an indicator related to income, which is used to evaluate the indicator of return on investment. The larger the average rate of return, the better. The alpha coefficient is an indicator related to income and is a relative index. The larger the alpha coefficient, the greater the ability of the fund to obtain excess returns. The standard deviation is a risk-related indicator, reflecting the fluctuation range of the fund recovery rate. The smaller the standard deviation, the better. Specifically, it refers to the deviation of the fund's monthly rate of return relative to the average monthly rate of return. The greater the monthly income fluctuation of the fund, the greater the corresponding standard deviation. The beta coefficient is a risk-related indicator used to measure price volatility, which is used to assess the volatility of a stock or a stock fund relative to the entire market. The bigger the beta coefficient is, the better the bull market or the rising phase, and the smaller the beta coefficient in the bear market or the down phase, the better. The Morningstar risk factor is a risk-related indicator used to calculate the risk of a downward float relative to a similar fund over a certain period of time. The greater the Morningstar risk indicator, the greater the risk of downward fluctuation. Therefore, the Morningstar risk coefficient is as small as possible. The Sharpe ratio is an indicator related to both income and risk, and is a standardized indicator of fund performance evaluation. The higher the Sharpe ratio, the better. R-square is a measure related to both income and risk, which is used to reflect the change of performance. R-square is a measure of the extent to which a change in the performance of a fund can be explained by the change of the benchmark index, from 0 to 100, the closer to 100 The more reliable the alpha and beta coefficients are.

In this embodiment, the fund types according to different risk levels may be recommended according to different fund evaluation indicators. For example, the fund products of the two types of funds, such as currency type and bond type, can be recommended based on two kinds of income-based fund evaluation indicators: average return rate and alpha coefficient; for the hybrid fund type, the fund product can be based on the Sharpe ratio. And R-square, two fund-based evaluation indicators based on income and risk; for fund products of both index type and stock type, there are three risk-based fund evaluations based on standard deviation, beta coefficient and morningstar risk coefficient. Indicators are recommended.

S43: According to the fund evaluation index, the quick-sorting algorithm is used to sort the recommended fund products, and the target fund products are determined.

Among them, the basic idea of the quick sort algorithm is to divide the data to be sorted into two independent parts by one sorting, in which all the data of one part is smaller than all the data of the other part, and then the two parts of data are used according to this method. Quick sorting is performed separately, and the entire sorting process can be performed recursively, so that the entire data becomes an ordered sequence. The fast sorting algorithm is the fastest algorithm in the internal sorting algorithm based on keyword comparison, and the algorithm is highly efficient.

In this embodiment, according to the fund evaluation index, the quick-sorting algorithm is used to sort the fund products to be recommended by the fund trading system, and the target fund products are obtained through the sorting result of the fund evaluation indexes corresponding to the fund products to be recommended, and the smart phone is obtained through the smart phone. The display interface of the terminal such as a tablet computer displays the target fund product. For example, taking the fund evaluation index as the average rate of return, using the quick sort algorithm to rank the average rate of return of the recommended fund products, the sorting result of the average rate of return can be obtained, and the highest rate of return among the ranked results corresponds to the fund to be recommended. Target fund products. Understandably, the target fund products viewed by the target users through terminals such as smart phones and tablet computers are sorted according to the fund evaluation indicators, so that the target users can understand the target fund products that match the investment conditions of the users, and may have It will help improve the accuracy of investors' purchase of fund products and reduce the risk of fund purchases.

In this embodiment, before the step S10 and the step S20, the fund product recommendation method further includes: obtaining a product recommendation instruction, so that the current user portrait data is acquired based on the product recommendation instruction in step S10, and the training is invoked based on the product recommendation instruction in step S20. User data model. In the fund trading system, the “fund recommendation” button may be displayed in the display interface, and the user may click the “fund recommendation” button to name the fund trading system to obtain the product recommendation instruction. Alternatively, in the fund trading system, the user may preset settings, and when the user logs into the fund trading system by using the pre-registered login account, the fund trading system may be triggered to obtain the product recommendation instruction, so that the display interface displays the target fund product.

In the fund product recommendation method provided by the embodiment, the target risk assessment value is determined based on the current user portrait data and the user data type, and the target target fund product is determined by using the target risk assessment value, so that the target fund recommended to the target user is obtained. The product accurately locates the user's own investment conditions and fund products, which helps the target users to improve the investment return rate.

It should be understood that the size of the serial number of each step in the above embodiments does not mean the order of execution order, and the order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the present application.

Example 2

Fig. 6 is a block diagram showing the principle of the fund product recommendation device corresponding to the fund product recommendation method in the first embodiment. As shown in FIG. 6, the fund product recommendation device includes a current user portrait data acquisition module 10, a user data model acquisition module 20, a target cluster cluster determination module 30, and a target fund product determination module 40. The implementation functions of the current user portrait data obtaining module 10, the user data model invoking module 20, the target clustering cluster determining module 30, and the target fund product determining module 40 correspond to the steps corresponding to the fund product recommendation method in the first embodiment. In order to avoid redundancy, the present embodiment will not be described in detail.

The current user portrait data obtaining module 10 is configured to acquire current user portrait data, and the current user portrait data includes at least one current feature data.

The user data model obtaining module 20 is configured to acquire a user data model, where the user data model includes at least two clustering clusters, and each clustering cluster corresponds to a risk assessment value.

The target clustering cluster determining module 30 is configured to acquire a target clustering cluster corresponding to the at least one current feature data from the at least two clustering clusters based on the current user image data.

The target fund product determining module 40 is configured to determine the target fund product based on the risk assessment value corresponding to the target clustering cluster.

Preferably, the fund product recommendation device further includes a user data model training module 50 and a user data model storage module 60.

The user data model training module 50 is configured to train the user data model based on the training user image data, and the training user image data includes at least one training feature data.

The user data model storage module 60 is configured to store the user data model in a database.

The user data model obtaining module 20 is configured to obtain a user data model from a database.

Preferably, the user data model training module 50 includes a normalization processing unit 51, a cluster cluster acquisition unit 52, a risk evaluation value acquisition unit 53, and a data model acquisition unit 54.

a normalization processing unit 51, configured to perform normalization processing on at least one standard feature data in the training user image data, so that the training user image data includes at least one standard feature data;

The clustering cluster acquiring unit 52 is configured to cluster at least one standard feature data in the training user portrait data by using a K-means clustering algorithm, and acquire at least two clustering clusters, each cluster cluster corresponding to a centroid User portrait data.

The risk evaluation value obtaining unit 53 is configured to perform weighting processing on the centroid user portrait data by using a weighting operation algorithm, and determine that the centroid user portrait data corresponds to a risk evaluation value; the weighting operation algorithm is

among them,

Pi is the risk assessment value of the centroid user, Vi is the value of each standard feature data in the centroid user portrait data, and Wi is the weight of each standard feature data.

The data model obtaining unit 54 is configured to acquire a user data model based on the clustering cluster and the risk assessment value.

Preferably, the target clustering class cluster determining module 30 includes an Euclidean distance acquiring unit 31 and a target clustering cluster selecting unit 32.

The Euclidean distance obtaining unit 31 is configured to calculate the current user image data and the centroid user image data of at least two cluster clusters in the user data model to obtain at least two Euclidean distances.

The target clustering cluster selection unit 32 is configured to select a cluster cluster in which the centroid user portrait data corresponding to the minimum value of the at least two Euclidean distances is located as the target cluster cluster corresponding to the at least one current feature data.

Preferably, the target fund product determining module 40 includes a fund type determining unit 41, an evaluation index obtaining unit 42, and a target fund product determining unit 43.

The fund type determining unit 41 is configured to determine a target fund type corresponding to the target risk assessment value based on the target risk assessment value.

The evaluation index obtaining unit 42 is configured to obtain the fund products to be recommended and the fund evaluation indicators corresponding to the fund type according to the fund type.

The target fund product determining unit 43 is configured to sort the recommended fund products by using a quick sorting algorithm according to the fund evaluation index, and determine the target fund product.

Example 3

The embodiment provides a computer readable storage medium having stored thereon computer readable instructions, which are implemented by the processor to implement the fund product recommendation method in Embodiment 1, in order to avoid duplication, here No longer. Alternatively, when the computer readable instructions are executed by the processor, the functions of the modules/units in the fund product recommendation device in Embodiment 2 are implemented. To avoid repetition, details are not described herein again.

Example 4

FIG. 7 is a schematic diagram of a terminal device according to an embodiment of the present application. As shown in FIG. 7, the terminal device 70 of this embodiment includes a processor 71, a memory 72, and computer readable instructions 73 stored in the memory 72 and operable on the processor 71, the processor 71 executing computer readable instructions Each step of the fund product recommendation method in Embodiment 1 is implemented at 73 hours, such as steps S10, S20, S30, and S40 shown in FIG. Alternatively, when the processor 71 executes the computer readable instructions 73, the functions of the modules/units of the fund product recommendation device in Embodiment 2 are implemented. As shown in FIG. 6, the current user profile data acquisition module 10, the user data model acquisition module 20, and the target aggregation The class cluster determination module 30 and the target fund product determination module 40 function.

Illustratively, computer readable instructions 73 may be partitioned into one or more modules/units, one or more modules/units being stored in memory 72 and executed by processor 71 to complete the application. The one or more modules/units may be an instruction segment of a series of computer readable instructions 73 capable of performing a particular function, which is used to describe the execution of computer readable instructions 73 in the terminal device 70. For example, the computer readable instructions 73 may be divided into a current user portrait data acquisition module 10, a user data model acquisition module 20, a target cluster cluster determination module 30, a target fund product determination module 40, a user data model training module 50, and a user. The data model storage module 60 has the following functions:

among them,

The terminal device 70 can be a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud server. The terminal device may include, but is not limited to, a processor 71, a memory 72. It will be understood by those skilled in the art that FIG. 7 is only an example of the terminal device 70, and does not constitute a limitation of the terminal device 70, and may include more or less components than those illustrated, or combine some components, or different components. For example, the terminal device may further include an input/output device, a network access device, a bus, and the like.

The processor 71 may be a central processing unit (CPU), or may be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc. The general purpose processor may be a microprocessor or the processor or any conventional processor or the like.

The memory 72 may be an internal storage unit of the terminal device 70, such as a hard disk or a memory of the terminal device 70. The memory 72 may also be an external storage device of the terminal device 70, such as a plug-in hard disk provided on the terminal device 70, a smart memory card (SMC), a Secure Digital (SD) card, and a flash memory card (Flash). Card) and so on. Further, the memory 72 may also include both an internal storage unit of the terminal device 70 and an external storage device. The memory 72 is used to store computer readable instructions 73 and other programs and data required by the terminal device. The memory 72 can also be used to temporarily store data that has been or will be output.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, only the division of each functional unit and module described above is exemplified. In practical applications, the above functions may be assigned to different functional units according to needs. The module is completed by dividing the internal structure of the device into different functional units or modules to perform all or part of the functions described above.

In addition, each functional unit in each embodiment of the present application 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 integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated modules/units, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, the present application implements all or part of the processes in the foregoing embodiments, and may also be implemented by computer readable instructions, which may be stored in a computer readable storage medium. The computer readable instructions, when executed by a processor, may implement the steps of the various method embodiments described above. Wherein, the computer readable instructions comprise computer readable instruction code, which may be in the form of source code, an object code form, an executable file or some intermediate form or the like. The computer readable medium can include any entity or device capable of carrying the computer readable instruction code, a recording medium, a USB flash drive, a removable hard drive, a magnetic disk, an optical disk, a computer memory, a read only memory (ROM, Read-Only) Memory), random access memory (RAM), electrical carrier signals, telecommunications signals, and software distribution media. It should be noted that the content contained in the computer readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in a jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, computer readable media It does not include electrical carrier signals and telecommunication signals.

The above-mentioned embodiments are only used to explain the technical solutions of the present application, and are not limited thereto; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still implement the foregoing embodiments. The technical solutions described in the examples are modified or equivalently replaced with some of the technical features; and the modifications or substitutions do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the present disclosure. Within the scope of protection of the application.

Claims

A fund product recommendation method, characterized in that it comprises:

Obtaining current user portrait data, the current user portrait data including at least one current feature data;

Obtaining a user data model, where the user data model includes at least two clustering clusters, each of the clustering clusters corresponding to a risk assessment value;

And acquiring, according to the current user portrait data, a target clustering cluster corresponding to at least one of the current feature data from at least two of the clustering clusters;

The target fund product is determined based on the risk assessment value corresponding to the target clustering cluster.
The method for recommending a fund product according to claim 1, wherein the method for recommending the fund product further comprises:

Training the user data model based on training user portrait data, the training user portrait data including at least one training feature data;

Storing the user data model in a database;

The acquiring the user data model includes: obtaining the user data model from the database.
The fund product recommendation method according to claim 2, wherein the training the user data model based on training user portrait data comprises:

Performing normalization processing on at least one of the training feature data in the training user image data, so that the training user image data includes at least one standard feature data;

The K-means clustering algorithm is used to cluster at least one of the standard feature data in the training user image data to obtain at least two clustering clusters, and each clustering cluster corresponds to a centroid user image data;

The weighting operation algorithm is used to perform weighting processing on the centroid user image data, and the centroid user image data is determined to correspond to a risk evaluation value, where the weighting operation algorithm is P i =∑V i ·W i , where ∑W i = 1, Pi is the risk assessment value of the centroid user, Vi is the value of each standard feature data in the centroid user portrait data, and Wi is the weight of each standard feature data;

The user data model is obtained based on the cluster cluster and the risk assessment value.
The fund product recommendation method according to claim 1, wherein the acquiring a target corresponding to at least one of the current feature data from at least two of the cluster clusters based on the current user portrait data Clustering clusters, including:

Calculating, respectively, the current user portrait data and the centroid user portrait data of at least two of the cluster clusters in the user data model to obtain at least two Euclidean distances;

And selecting a cluster cluster in which the centroid user image data corresponding to the minimum value of the at least two of the Euclidean distances is located as the target cluster cluster corresponding to the at least one of the current feature data.
The fund product recommendation method according to claim 1, wherein the determining the target fund product based on the risk assessment value corresponding to the target clustering cluster comprises:

Determining a corresponding target fund type based on the risk assessment value corresponding to the target clustering cluster;

Obtaining a fund product and fund evaluation index corresponding to the fund type according to the type of the fund;

According to the fund evaluation index, the quick-sorting algorithm is used to sort the products to be recommended to determine the target fund product.
A fund product recommendation device, comprising:

a current user image data obtaining module, configured to acquire current user portrait data, where the current user portrait data includes at least one current feature data;

a user data model obtaining module, configured to acquire a user data model, where the user data model includes at least two clustering clusters, each of the clustering clusters corresponding to a risk assessment value;

a target clustering cluster determining module, configured to acquire, according to the current user portrait data, a target clustering cluster corresponding to at least one of the current feature data from at least two of the clustering clusters;

The target fund product determining module is configured to determine the target fund product based on the risk assessment value corresponding to the target clustering cluster.
The fund product recommendation device according to claim 6, wherein the fund product recommendation device further comprises:

a user data model training module, configured to train the user data model based on training user portrait data, the training user portrait data including at least one training feature data;

a user data model storage module, configured to store the user data model in a database;

The user data model obtaining module is configured to obtain the user data model from the database.
The fund product recommendation device according to claim 7, wherein the user data model training module comprises:

a normalization processing unit, configured to perform normalization processing on at least one of the standard feature data in the training user image data, so that the training user image data includes at least one standard feature data;

a clustering cluster acquiring unit, configured to cluster at least one of the standard feature data in the training user portrait data by using a K-means clustering algorithm, to acquire at least two clustering clusters, each cluster cluster Corresponding to a centroid user image data;

a risk evaluation value obtaining unit, configured to perform weighting processing on the centroid user image data by using a weighting operation algorithm, and determine that the centroid user image data corresponds to a risk evaluation value;

And a data model obtaining unit, configured to acquire the user data model based on the clustering cluster and the risk assessment value.
The fund product recommendation device according to claim 6, wherein the target clustering cluster determining module comprises:

An Euclidean distance acquisition unit, configured to calculate the current user image data and the centroid user image data of at least two of the cluster clusters in the user data model to obtain at least two Euclidean distances;

a target clustering cluster selecting unit, configured to select a clustering cluster in which at least two of the at least two of the Euclidean distances correspond to the centroid user image data as the target cluster corresponding to the at least one of the current feature data Class class cluster.
The fund product recommendation device according to claim 6, wherein the target fund product determination module comprises:

a fund type determining unit, configured to determine a target fund type corresponding to the target risk assessment value based on the target risk assessment value;

An evaluation index obtaining unit, configured to obtain, according to the fund type, a fund product and fund evaluation index corresponding to the fund type;

The target fund product determining unit is configured to sort the to-be-recommended fund products by using a quick sorting algorithm according to the fund evaluation index, and determine the target fund product.
A terminal device comprising a memory, a processor, and computer readable instructions stored in the memory and operable on the processor, wherein the processor executes the computer readable instructions as follows step:

Obtaining current user portrait data, the current user portrait data including at least one current feature data;

Obtaining a user data model, where the user data model includes at least two clustering clusters, each of the clustering clusters corresponding to a risk assessment value;

And acquiring, according to the current user portrait data, a target clustering cluster corresponding to at least one of the current feature data from at least two of the clustering clusters;

The target fund product is determined based on the risk assessment value corresponding to the target clustering cluster.
The terminal device according to claim 11, wherein before said step of acquiring a user data model, said processor further implements the following steps when said computer readable instructions are executed:

Training the user data model based on training user portrait data, the training user portrait data including at least one training feature data;

Storing the user data model in a database;

The acquiring the user data model includes: obtaining the user data model from the database.
The terminal device according to claim 12, wherein the training the user data model based on training user portrait data comprises:

Performing normalization processing on at least one of the training feature data in the training user image data, so that the training user image data includes at least one standard feature data;

The K-means clustering algorithm is used to cluster at least one of the standard feature data in the training user image data to obtain at least two clustering clusters, and each clustering cluster corresponds to a centroid user image data;

The weighting operation algorithm is used to perform weighting processing on the centroid user image data, and the centroid user image data is determined to correspond to a risk evaluation value, where the weighting operation algorithm is P i =∑V i ·W i , where ∑W i = 1, Pi is the risk assessment value of the centroid user, Vi is the value of each standard feature data in the centroid user portrait data, and Wi is the weight of each standard feature data;

The user data model is obtained based on the cluster cluster and the risk assessment value.
The terminal device according to claim 11, wherein the acquiring, based on the current user portrait data, target clusters corresponding to at least one of the current feature data from at least two cluster clusters Class clusters, including:

Calculating, by the current user image data, the centroid user image data of at least two of the cluster clusters in the user data model, to obtain at least two Euclidean distances;

And selecting a cluster cluster in which the centroid user image data corresponding to the minimum value of the at least two of the Euclidean distances is located as the target cluster cluster corresponding to the at least one of the current feature data.
The terminal device according to claim 11, wherein the determining the target fund product based on the risk assessment value corresponding to the target clustering cluster comprises:

Determining a corresponding target fund type based on the risk assessment value corresponding to the target clustering cluster;

Obtaining a fund product and fund evaluation index corresponding to the fund type according to the type of the fund;

According to the fund evaluation index, the quick-sorting algorithm is used to sort the products to be recommended to determine the target fund product.
A computer readable storage medium storing computer readable instructions, wherein the computer readable instructions, when executed by a processor, implement the following steps:

Obtaining current user portrait data, the current user portrait data including at least one current feature data;

Obtaining a user data model, where the user data model includes at least two clustering clusters, each of the clustering clusters corresponding to a risk assessment value;

And acquiring, according to the current user portrait data, a target clustering cluster corresponding to at least one of the current feature data from at least two of the clustering clusters;

The target fund product is determined based on the risk assessment value corresponding to the target clustering cluster.
The computer readable storage medium of claim 16 wherein said computer readable instructions, when executed by said processor, further implement the following steps prior to said step of obtaining a user data model:

Training the user data model based on training user portrait data, the training user portrait data including at least one training feature data;

Storing the user data model in a database;

The acquiring the user data model includes: obtaining the user data model from the database.
The computer readable storage medium of claim 17, wherein the training the user data model based on training user portrait data comprises:

Performing normalization processing on at least one of the training feature data in the training user image data, so that the training user image data includes at least one standard feature data;

The K-means clustering algorithm is used to cluster at least one of the standard feature data in the training user image data to obtain at least two clustering clusters, and each clustering cluster corresponds to a centroid user image data;

The weighting operation algorithm is used to perform weighting processing on the centroid user image data, and the centroid user image data is determined to correspond to a risk evaluation value, where the weighting operation algorithm is P i =∑V i ·W i , where ∑W i = 1, Pi is the risk assessment value of the centroid user, Vi is the value of each standard feature data in the centroid user portrait data, and Wi is the weight of each standard feature data;

The user data model is obtained based on the cluster cluster and the risk assessment value.
The computer readable storage medium according to claim 16, wherein said acquiring, based on said current user portrait data, at least two of said clustering clusters corresponds to at least one of said current feature data Target clustering clusters, including:

Calculating, by the current user image data, the centroid user image data of at least two of the cluster clusters in the user data model, to obtain at least two Euclidean distances;

And selecting a cluster cluster in which the centroid user image data corresponding to the minimum value of the at least two of the Euclidean distances is located as the target cluster cluster corresponding to the at least one of the current feature data.
The computer readable storage medium according to claim 16, wherein the determining the target fund product based on the risk assessment value corresponding to the target clustering cluster comprises:

Determining a corresponding target fund type based on the risk assessment value corresponding to the target clustering cluster;

According to the fund type, obtaining a fund product and fund evaluation index corresponding to the fund type; according to the fund evaluation index, using a quick sorting algorithm to sort the fund products to be recommended, and determining the target fund product.