Recommendation method, system and device for investment and storage medium

Abstract

The invention relates to the technical field of data processing, and discloses an investment recommendation method, system and device and a storage medium, and the method comprises the steps: obtaining fund product information and user behavior characteristics, building a fund product information database and a user behavior characteristic database, carrying out the first weight calculation according to the databases, and obtaining a first weight; performing second weight calculation according to the matching attribute weight to obtain feature vectorized data, performing first matching degree calculation according to the feature vectorized data to obtain a first matching degree, and performing second matching degree calculation based on support vector machine classification according to a user behavior feature database to obtain a second matching degree; and obtaining a second matching degree, performing product calculation according to the first matching degree and the second matching degree to obtain a third matching degree, and recommending the fund product to the user by taking the third matching degree as a standard. The method has the effect of recommending fund products with matched information and high accuracy to the user.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a recommendation method, system, device and storage medium for investment. Background Art

[0002] At present, with the rapid development of Internet finance, data processing technology has been widely used in the financial field. Data processing provides investors with convenient asset allocation services through technical means such as big data analysis, cloud computing and personalized recommendations.

[0003] In one prior art, when pushing fund product related information to users, it is often impossible to accurately analyze the matching degree between users and fund products, and it is impossible to meet the personalized investment needs of users.

[0004] In the prior art, there are problems in that the fund product information recommended to users does not match and the accuracy of the recommended fund product information is low. Summary of the invention

[0005] The present invention provides a recommendation method, system, device and storage medium for investment, so as to solve the problem that the fund product information recommended to the user does not match and the accuracy of the recommended fund product information is low.

[0006] In a first aspect, in order to solve the above technical problems, the present invention provides a recommendation method for investment, comprising:

[0007] Obtain fund product information and user behavior characteristics, and establish a fund product information database and a user behavior characteristics database;

[0008] Performing a first weight calculation based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund;

[0009] Perform a second weight calculation based on the matching attribute weights to obtain quantized data of user behavior characteristics and fund product characteristics;

[0010] Performing a first matching degree calculation based on the feature vector quantized data to obtain a first matching degree; wherein the first matching degree is a matching degree between the fund product information database and the user behavior feature database;

[0011] According to the user behavior feature database, based on support vector machine classification, a second matching degree calculation is performed to obtain a second matching degree, wherein the second matching degree is a matching degree between the user behavior feature and the fund product information;

[0012] A product calculation is performed based on the first matching degree and the second matching degree to obtain a third matching degree, and the fund product is recommended to the user based on the third matching degree as the standard matching degree.

[0013] In an optional implementation, the fund product information database includes project code, project name, manager code;

[0014] The user behavior feature database includes the number of all fund products browsed by the user, the code of the fund product most recently browsed by the user, the time the user browsed a single fund product, the total time the user browsed fund products, the codes of the fund products in which the user invested, the fund products purchased by the user and the amount of investment products purchased by the user.

[0015] In an optional implementation, performing a first weight calculation based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund includes:

[0016] The formula for calculating the first weight is:

[0017] W1=w×u×u2

[0018] Among them, W1 is the matching attribute weight between users and funds, w is the time weight of users browsing products, u is the number weight of funds, and u2 is the investment amount weight;

[0019] Where w = T / T0, u = V / U0, u2 = a × A;

[0020] Among them, T is the time the user browses a single fund product, T0 is the total time the user browses fund products, V is the number of all fund products browsed by the user, U0 is the total number of all fund products in the fund product information database, a is the preset amount ratio coefficient used to standardize the amount value, and A is the amount of the user purchasing the investment product.

[0021] In an optional implementation, performing a second weight calculation according to the matching attribute weight to obtain quantized data of user behavior characteristics and fund product characteristics includes:

[0022] A formula is calculated based on the matching attribute weight and the time weight of the user browsing the product to obtain the time feature weight of the user browsing the product;

[0023] According to the time feature weight, the fund products are classified based on the support vector machine to obtain various categories of fund products, and the standard deviation of each category is calculated to obtain user behavior characteristics and feature vector quantized data of fund products.

[0024] In an optional implementation, performing a first matching degree calculation based on the feature vectorized data to obtain the first matching degree includes:

[0025] The first matching degree calculation is a cosine similarity calculation, and the specific formula is:

[0026]

[0027] Among them, P is the matching degree between the fund product information database and the user behavior feature database, A is the feature vector quantized data of the fund product, and B is the feature vector quantized data of the user behavior feature.

[0028] In an optional implementation, performing a second matching degree calculation based on the user behavior feature database and support vector machine classification to obtain the second matching degree includes:

[0029] The formula for calculating the second matching degree is:

[0030]

[0031] Among them, Q is the matching degree between user behavior characteristics and fund product information, x is user behavior characteristics, and x i is the fund product information, n is the number of support vectors, y i is the label of the i-th support vector, α i is the Lagrange multiplier of the i-th training sample, which represents the weight of the i-th training sample in the final decision function. b is the bias term of the model, which allows the decision boundary to move in the feature space. K(x,x i ) is a kernel function, which is used to calculate the similarity between user behavior characteristics and fund product attributes to determine the degree of match between them.

[0032] In an optional implementation, performing product calculation according to the first matching degree and the second matching degree to obtain a third matching degree, and using the third matching degree as a standard matching degree to recommend a fund product to a user includes:

[0033] Sorting the fund products in the third matching degree from large to small, and generating a recommendation list, and recommending the fund products in the recommendation list to the user in order from large to small;

[0034] Wherein, the formula for calculating the product is:

[0035] S=P×Q

[0036] Among them, S is the third matching degree, Q is the second matching degree, that is, the matching degree between the user behavior characteristics and the fund product information, and P is the first matching degree, that is, the matching degree between the fund product information database and the user behavior characteristics database.

[0037] In a second aspect, the present invention provides an investment dynamic recommendation system based on matching analysis, comprising:

[0038] The data acquisition module is used to acquire fund product information and user behavior characteristics, and establish a fund product information database and a user behavior characteristics database;

[0039] An attribute weight module, used to perform a first weight calculation based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund;

[0040] A feature vectorization module, used to perform a second weight calculation according to the matching attribute weight to obtain feature vectorization data of user behavior features and fund products;

[0041] A fund product matching module, configured to perform a first matching degree calculation based on the feature vector quantized data to obtain a first matching degree, wherein the first matching degree is a matching degree between the fund product information database and the user behavior feature database;

[0042] A user behavior feature matching module, configured to perform a second matching degree calculation based on the user behavior feature database and support vector machine classification to obtain a second matching degree, wherein the second matching degree is a matching degree between the user behavior feature and the fund product information;

[0043] The product recommendation module is used to perform product calculation based on the first matching degree and the second matching degree to obtain a third matching degree, and recommend fund products to users based on the third matching degree as a standard matching degree.

[0044] In a third aspect, the present invention further provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements any one of the above-mentioned recommendation methods for investment.

[0045] In a fourth aspect, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned recommendation methods for investment.

[0046] The present invention provides a recommendation method, system, device and storage medium for investment. The solution establishes a fund product information database and a user behavior feature database, and uses a support vector machine to match and analyze user behavior features with fund product attributes to achieve accurate recommendation of investment dynamics. Specifically, the method includes obtaining fund product information and user behavior features, performing weight calculation, obtaining the matching attribute weight of the user and the fund, and then finally recommending fund products that match the user's investment preference through feature quantization and matching degree calculation.

[0047] In the derivation process, the present invention first establishes a fund product information database and a user behavior feature database by acquiring fund product information and user behavior features; then, based on the fund product information database and the user behavior feature database, a first weight calculation is performed to obtain the matching attribute weight between the user and the fund; then, based on the matching attribute weight, a second weight calculation is performed to obtain the feature vector quantized data of the user behavior features and the fund product; then, based on the feature vector quantized data, a first matching degree calculation is performed to obtain a first matching degree; wherein, the first matching degree is the matching degree between the fund product information database and the user behavior feature database; then, based on the user behavior feature database, a second matching degree calculation is performed based on support vector machine classification to obtain a second matching degree, wherein, the second matching degree is the matching degree between the user behavior features and the fund product information; then, based on the first matching degree and the second matching degree, a product calculation is performed to obtain a third matching degree, and the third matching degree is used as the standard matching degree to recommend the fund product to the user.

[0048] Compared with the prior art, the present invention has the following beneficial effects: First, it solves the problem of low accuracy of information recommendation in the prior art, and improves the accuracy of investment information push by accurately matching user behavior characteristics and fund product attributes. Secondly, the method can quickly recommend suitable fund products according to the user's personal investment preferences, shorten the user's purchase decision time, and improve purchase efficiency. Finally, the present invention also reduces the sales cost of the product, improves user satisfaction through personalized recommendations, realizes customized push services for financial information, and meets the personalized needs of different users. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a flowchart of a recommendation method for investment provided by the first embodiment of the present invention;

[0050] Figure 2 It is a schematic diagram of the structure of an investment dynamic recommendation system based on matching analysis provided by the second embodiment of the present invention. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] At present, with the rapid development of Internet finance, data processing technology has been widely used in the financial field. Data processing provides investors with convenient asset allocation services through technical means such as big data analysis, cloud computing and personalized recommendations.

[0053] In one prior art, when pushing fund product related information to users, it is often impossible to accurately analyze the degree of match between users and fund products, and it is impossible to meet the personalized investment needs of users. This shows that there are problems in the prior art of mismatched fund product information recommended to users and low accuracy of recommended fund product information.

[0054] In order to solve the above technical problems, refer to Figure 1 The first embodiment of the present invention provides a recommendation method for investment, comprising the following steps:

[0055] S11, obtaining fund product information and user behavior characteristics, and establishing a fund product information database and a user behavior characteristics database;

[0056] S12, performing a first weight calculation based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund;

[0057] S13, performing a second weight calculation according to the matching attribute weight to obtain quantized data of user behavior characteristics and fund product characteristics;

[0058] S14, performing a first matching degree calculation based on the feature vector quantized data to obtain a first matching degree; wherein the first matching degree is a matching degree between the fund product information database and the user behavior feature database;

[0059] S15, performing a second matching degree calculation based on the support vector machine classification according to the user behavior feature database to obtain a second matching degree, wherein the second matching degree is a matching degree between the user behavior feature and the fund product information;

[0060] S16, performing a product calculation based on the first matching degree and the second matching degree to obtain a third matching degree, and using the third matching degree as a standard matching degree to recommend the fund product to the user.

[0061] In step S11, fund product information and user behavior characteristics are obtained, and a fund product information database and a user behavior characteristic database are established;

[0062] According to the steps of acquiring fund product information and user behavior characteristics, a fund product information database and a user behavior characteristic database are established. A key feature of the steps is that two core databases are constructed: a fund product information database and a user behavior characteristic database.

[0063] The fund product information database includes detailed attributes of fund products, such as project code, project name, manager code, etc. These attributes provide a comprehensive description of the fund products.

[0064] The user behavior feature database records the behavior patterns of users interacting with fund products, including the number of all fund products browsed by the user, the code of the fund product most recently browsed by the user, the time the user browsed a single fund product, the total time the user browsed fund products, the code of the fund products the user has invested in, the fund products purchased by the user, and the amount of investment products purchased by the user. These features together form a detailed portrait of user preferences and behavior patterns.

[0065] The acquisition of features in this step is achieved through automated data collection technology. For fund product information, it can be imported from databases provided by financial institutions, public financial reports, market data publishing platforms and other sources through crawler technology or API interfaces. User behavior features are automatically captured through user interactions on the investment platform, such as browsing records, purchase history and search habits. These data are pre-processed, including cleaning, denoising and formatting, to ensure data quality and consistency, and then stored in the corresponding database to provide support for further analysis and application.

[0066] In this step, the fund product information and user behavior characteristics obtained directly serve to build a rich database resource. These databases not only provide a basis for subsequent data analysis and processing, but also enable the system to have a comprehensive and in-depth understanding of fund products and user behavior. Through these characteristics, the system can capture market dynamics and changes in user needs, thereby establishing an accurate matching framework between users and fund products. This matching framework is the basis for realizing personalized investment recommendations. It allows the system to recommend investment products that best meet user needs based on the user's specific behavior and preferences, as well as the detailed attributes of the fund product.

[0067] In step S12, a first weight calculation is performed based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund;

[0068] An innovative feature of this step is to use the fund product information database and the user behavior feature database for comprehensive analysis, and calculate the matching attribute weights between users and fund products through algorithms. This process involves quantifying the user's behavioral preferences for fund products and the multi-dimensional features of the products, and converting these quantified results into weights. The weights can well reflect the user's preference for different fund product features, as well as the strength of the association between fund product features and user behavior. These weights are the basis for subsequent matching analysis and recommendation decisions.

[0069] The acquisition of the matching weight is achieved through an algorithm model, specifically, it is determined by the first weight calculation formula. This formula takes into account the time (T) that the user browses a single fund product, the total time T0 that the user browses the fund product, the number of all fund products browsed by the user (V), the total number of all fund products in the fund product information database U0, and the amount (A) that the user purchases the investment product, combined with the preset amount ratio coefficient a. These parameters reflect the user's attention and participation in the fund product, as well as the user's investment behavior and preferences. In this way, the algorithm can quantify the user's emphasis on each feature of the fund product, thereby scoring the degree of matching between each user and the fund product feature.

[0070] In this step, the weight features obtained are used to evaluate the initial match between the user and the fund product. These weights, as quantitative indicators, reflect the relative preference strength of the user for each attribute of the fund product. For example, if the user tends to browse fund products with a specific manager or company, the corresponding weight will be higher, indicating that the user prefers these features. These weights are directly used to calculate the match between the user and the fund product, providing a basis for investment recommendations. However, it should be noted that these weights are not directly used to generate the final recommendation list, but to provide input data for subsequent match calculations. In this step, the weights are only used to evaluate and quantify the initial match between the user and the fund product, and are not involved in the subsequent sorting or recommendation decision process.

[0071] The formula for calculating the first weight is:

[0072] W1=w×u×u2

[0073] Among them, W1 is the matching attribute weight between users and funds, w is the time weight of users browsing products, u is the number weight of funds, and u2 is the investment amount weight;

[0074] Where w = T / T0, u = V / U0, u2 = a × A;

[0075] Among them, T is the time the user browses a single fund product, T0 is the total time the user browses fund products, V is the number of all fund products browsed by the user, U0 is the total number of all fund products in the fund product information database, a is the preset amount ratio coefficient used to standardize the amount value, and A is the amount of the user purchasing the investment product.

[0076] Among them, a can be set to a very small value such as 0.000001, so that the amount can become a smaller value to facilitate statistical calculations.

[0077] In step S13, a second weight calculation is performed according to the matching attribute weight to obtain quantized data of user behavior characteristics and fund product characteristics;

[0078] The core feature of this step is to quantify the impact of various attributes of fund products and user behavior characteristics, and generate quantized data of fund product characteristics. This process comprehensively considers multiple attribute dimensions of fund products, such as yield, risk level, investment period, etc., and combines the user's preference weights for these attributes, and uses mathematical methods to convert these multidimensional attributes into a numerical vector. The weight of each dimension reflects the importance of the attribute in the overall evaluation, and the quantized data is convenient for subsequent calculations and processing.

[0079] The acquisition of feature vector quantization data in this step is achieved through the second weight calculation. This process is based on the obtained matching attribute weights, taking into account factors such as the time weight of the user browsing the product, the number weight of the fund, the investment amount weight, etc., combined with various attributes of the fund product, such as historical yield, risk level, etc., and the weighted value of each attribute is calculated through mathematical formulas. These weighted values ​​comprehensively reflect user preferences and fund product attributes. Through specific algorithms, such as support vector machines or other machine learning methods, these weighted attributes are converted into vector form. Each dimension represents an attribute, and the value in the vector represents the weighted contribution of the attribute.

[0080] In this step, the quantized data of the feature vectors of fund products are used to evaluate the similarities or differences between fund products. By calculating the distance between the feature vectors of different fund products (such as Euclidean distance, cosine similarity, etc.), it is possible to determine which fund products are closer in attributes and which products are farther apart. This process does not directly involve user information, but focuses on the feature analysis of the fund products themselves. The obtained quantized data of the feature vectors provides an objective and standardized method for the comparison and selection of fund products, which helps to make more accurate matching analysis and recommendations in subsequent steps. For example, fund products can be clustered based on feature vector data, similar products can be grouped together, or the fund products that best match the user's preferences can be quickly screened out based on the user's behavioral characteristics. This method improves the efficiency and accuracy of the recommendation system, enabling it to better meet the personalized needs of users.

[0081] In step S14, a first matching degree calculation is performed based on the feature vector quantized data to obtain a first matching degree; wherein the first matching degree is the matching degree between the fund product information database and the user behavior feature database;

[0082] This step is characterized by using the feature vector quantization data to calculate the matching degree between the fund product and the user, that is, the first matching degree. This matching degree is obtained by comparing the quantitative characteristics of the fund product with the behavioral characteristics of the user, reflecting the fit between the characteristics of the fund product and the user's preferences. This quantitative method can objectively evaluate the attractiveness of the fund product to a specific user and provide an accurate basis for personalized recommendations.

[0083] In this step, the obtained feature vector quantization data is directly used to calculate the first matching degree between the fund product and the user. This calculation is done by comparing the feature vector of the fund product with the behavioral feature vector of the user, and usually involves calculating the similarity or distance between the two vectors. For example, cosine similarity can be used to measure the direction difference between the two vectors, or Euclidean distance can be used to measure the length difference between them. These calculation results provide a quantitative indicator, namely the first matching degree, which reflects the closeness of the match between the fund product characteristics and the user preferences.

[0084] The first matching degree calculation is a cosine similarity calculation, and the specific formula is:

[0085]

[0086] Among them, P is the matching degree between the fund product information database and the user behavior feature database, A is the feature vector quantized data of the fund product, and B is the feature vector quantized data of the user behavior feature.

[0087] Among them, A and B are two vectors to be compared, A·B represents the dot product of the vectors, and ||A|| and ||B|| represent the modulus of the vectors. This value ranges from -1 to 1, and the closer the value is to 1, the higher the degree of match.

[0088] In step S15, a second matching degree calculation is performed based on the support vector machine classification according to the user behavior feature database to obtain a second matching degree, wherein the second matching degree is a matching degree between the user behavior feature and the fund product information;

[0089] This step is characterized by applying the supervised learning technology of support vector machine (SVM) to classify and quantify the matching degree between user behavior characteristics and fund product information based on the data in the user behavior characteristic database, thereby calculating the second matching degree. This matching degree reflects the correspondence between the user's personal investment preferences and the characteristics of the fund product, and provides a quantitative indicator to measure the accuracy of the matching between the user and the fund product. Through the SVM classifier, this step can effectively distinguish and evaluate the user's preferences for different fund products, thereby providing relevant decision support for personalized recommendations.

[0090] The acquisition of features in this step relies on the user behavior feature database, and the data in the database is used to train the support vector machine model so that it can identify the relationship between user behavior patterns and fund product characteristics. Through this process, SVM can learn the complex boundaries of user preferences for fund products and classify user preferences accordingly.

[0091] In this step, the obtained user behavior features are directly used in the support vector machine model to calculate the second matching degree between the user behavior features and the fund product information. This calculation is completed by inputting the user's behavior data into the trained SVM model, and the model will output a numerical value indicating the matching degree between the user and the fund product. This numerical value can be regarded as a scoring system, and the higher the score, the higher the matching degree between the user and the fund product. This scoring is independent of other steps. It only reflects the matching results between the user's behavior features and the fund product information in the SVM model, and does not involve how to use these matching degrees to make the final fund product recommendation. Therefore, the output of this step provides an important input for the subsequent recommendation process, but does not determine the final recommendation output.

[0092] The formula for calculating the second matching degree is:

[0093]

[0094] Among them, Q is the matching degree between user behavior characteristics and fund product information, x is user behavior characteristics, and x iis the fund product information, n is the number of support vectors, y i is the label of the i-th support vector, α i is the Lagrange multiplier of the i-th training sample, which represents the weight of the i-th training sample in the final decision function. b is the bias term of the model, which allows the decision boundary to move in the feature space. K(x,x i ) is a kernel function, which is used to calculate the similarity between user behavior characteristics and fund product attributes to determine the degree of match between them.

[0095] Among them, Q is the output function of the support vector machine model, which represents the value of the input sample x in the decision function.

[0096] is the kernel function part in the support vector machine, where n is the number of training samples, y i is the label of the i-th support vector, α i is the Lagrange multiplier of the i-th training sample, K(x,x i ) is a kernel function, which calculates the similarity between the input sample x and the i-th training sample. The kernel function can be a linear kernel, a polynomial kernel, a radial basis function (RBF) kernel, etc.

[0097] α i : is the Lagrange multiplier of the ith training sample, which represents the weight of the ith training sample in the final decision function. During the training process of the support vector machine, these coefficients are determined by solving the optimization problem.

[0098] K(x,x i ): is a kernel function, which is a symmetric function used to calculate the inner product of two vectors in the feature space without explicitly mapping the vectors to a high-dimensional space. Exemplarily, kernel functions include linear kernels, polynomial kernels, and radial basis function (RBF) kernels.

[0099] b: is the bias term of the model, which allows the decision boundary to move in the feature space. During the training process of the support vector machine, the bias term b is also determined through the optimization problem.

[0100] y i : is the label of the i-th support vector, which is +1 or -1 for the binary classification problem.

[0101] In step S16, a product calculation is performed based on the first matching degree and the second matching degree to obtain a third matching degree, and the fund product is recommended to the user based on the third matching degree as the standard matching degree.

[0102] This step is characterized by combining two independent matching degrees, namely the first matching degree and the second matching degree, and calculating a comprehensive third matching degree based on the first matching degree and the second matching degree. This third matching degree is a weighted quantitative indicator that combines the direct matching degree between the fund product and the user's preference (the first matching degree) and the matching degree between the user's behavioral characteristics and the fund product information (the second matching degree). This comprehensive indicator is intended to provide a more comprehensive assessment to reflect the attractiveness and suitability of the fund product for a specific user.

[0103] The third matching degree feature is obtained by multiplying two independent matching degrees. Specifically, the first matching degree reflects the direct correlation between fund product attributes and user preferences, while the second matching degree is based on the matching analysis of user behavior and fund product information. The calculation of these two matching degrees is based on the technical methods such as support vector machine classification and feature vectorization used in the previous steps. By setting a product calculation formula, these two matching degrees are combined to reflect the comprehensive matching degree between users and fund products.

[0104] In this step, the third matching degree is used as the final recommendation basis. According to the calculated third matching degree, the system sorts all fund products from high to low according to the matching score. According to the sorted list, the fund products with the highest matching degree are selected, a recommendation list is generated, and these products are recommended to the user. This recommendation list not only takes into account the user's personal investment preferences, but also takes into account the various attributes of the fund products, thereby achieving highly personalized investment recommendations.

[0105] The product calculation formula in this step is:

[0106] S=P×Q

[0107] Among them, S is the third matching degree, Q is the second matching degree, that is, the matching degree between the user behavior characteristics and the fund product information, and P is the first matching degree, that is, the matching degree between the fund product information database and the user behavior characteristics database.

[0108] In summary, the present invention relates to the field of data processing technology, and discloses a recommendation method, system, device and storage medium for investment, the method comprising obtaining fund product information and user behavior characteristics, establishing a fund product information database and a user behavior characteristic database, performing a first weight calculation based on the database to obtain a matching attribute weight between the user and the fund, performing a second weight calculation based on the matching attribute weight to obtain feature vector quantized data of the user behavior characteristics and the fund product, performing a first matching degree calculation based on the feature vector quantized data to obtain a first matching degree, performing a second matching degree calculation based on the user behavior characteristic database and support vector machine classification to obtain a second matching degree, performing a product calculation based on the first and second matching degrees to obtain a third matching degree, and recommending the fund product to the user based on the third matching degree.

[0109] The existing technology has the problem of low accuracy of information recommendation and inability to meet the personalized investment needs of users. In traditional data processing services, although asset allocation services are provided to investors through technical means such as big data analysis and cloud computing, after users make investment decisions, the information related to fund products pushed often has information mismatch problems. This is because there is a serious conflict between the massive amount of financial data and the limited investment time of users, resulting in a large amount of product attribute information being displayed to users, which will cause investment decisions to fall into information overload, affect users' understanding of products, and thus fail to accurately push truly matching information, and the investment experience cannot meet expectations.

[0110] Compared with the prior art, the advantage of the present invention is that it realizes accurate recommendation of investment dynamics by comprehensively considering the matching degree between user behavior characteristics and fund product attributes. The present invention uses machine learning technologies such as support vector machines to perform feature vectorization and weight calculation on user behavior characteristics and fund product attributes, thereby accurately analyzing the matching degree between users and fund products. This method not only improves the accuracy of investment information push, shortens the user's purchase decision time, but also reduces the sales cost of the product, improves user satisfaction and purchase efficiency. In addition, the present invention can quickly recommend suitable fund products according to the user's personal investment preferences, realizes customized push services for financial information, and meets the personalized needs of different users.

[0111] Reference Figure 2 The second embodiment of the present invention provides an investment dynamic recommendation system based on matching analysis, comprising:

[0112] The data acquisition module is used to acquire fund product information and user behavior characteristics, and establish a fund product information database and a user behavior characteristics database;

[0113] An attribute weight module, used to perform a first weight calculation based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund;

[0114] A feature vectorization module, used to perform a second weight calculation according to the matching attribute weight to obtain feature vectorization data of user behavior features and fund products;

[0115] A fund product matching module, configured to calculate a first matching degree according to the feature vector quantized data to obtain a first matching degree, wherein the first matching degree is a matching degree between the fund product information database and the user behavior feature database;

[0116] A user behavior feature matching module, configured to perform a second matching degree calculation based on the user behavior feature database and support vector machine classification to obtain a second matching degree, wherein the second matching degree is a matching degree between the user behavior feature and the fund product information;

[0117] The product recommendation module is used to perform product calculation based on the first matching degree and the second matching degree to obtain a third matching degree, and recommend fund products to users based on the third matching degree as a standard matching degree.

[0118] The fund product information database includes project code, project name, and manager code;

[0119] The user behavior feature database includes the number of all fund products browsed by the user, the code of the fund product most recently browsed by the user, the time the user browsed a single fund product, the total time the user browsed fund products, the codes of the fund products in which the user invested, the fund products purchased by the user and the amount of investment products purchased by the user.

[0120] In an optional implementation, performing a first weight calculation based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund includes:

[0121] The formula for calculating the first weight is:

[0122] W1=w×u×u2

[0123] Among them, W1 is the matching attribute weight between users and funds, w is the time weight of users browsing products, u is the number weight of funds, and u2 is the investment amount weight;

[0124] Where w = T / T0, u = V / U0, u2 = a × A;

[0125] Among them, T is the time the user browses a single fund product, T0 is the total time the user browses fund products, V is the number of all fund products browsed by the user, U0 is the total number of all fund products in the fund product information database, a is the preset amount ratio coefficient used to standardize the amount value, and A is the amount of the user purchasing the investment product.

[0126] In an optional implementation, performing a second weight calculation according to the matching attribute weight to obtain quantized data of user behavior characteristics and fund product characteristics includes:

[0127] A formula is calculated based on the matching attribute weight and the time weight of the user browsing the product to obtain the time feature weight of the user browsing the product;

[0128] According to the time feature weight, the fund products are classified based on the support vector machine to obtain various categories of fund products, and the standard deviation of each category is calculated to obtain user behavior characteristics and feature vector quantized data of fund products.

[0129] In an optional implementation, performing a first matching degree calculation based on the feature vectorized data to obtain the first matching degree includes:

[0130] The first matching degree calculation is a cosine similarity calculation, and the specific formula is:

[0131]

[0132] Among them, P is the matching degree between the fund product information database and the user behavior feature database, A is the feature vector quantized data of the fund product, and B is the feature vector quantized data of the user behavior feature.

[0133] In an optional implementation, performing a second matching degree calculation based on the user behavior feature database and support vector machine classification to obtain the second matching degree includes:

[0134] The formula for calculating the second matching degree is:

[0135]

[0136] Among them, Q is the matching degree between user behavior characteristics and fund product information, x is user behavior characteristics, and x i is the fund product information, n is the number of support vectors, y i is the label of the i-th support vector, α i is the Lagrange multiplier of the i-th training sample, which represents the weight of the i-th training sample in the final decision function. b is the bias term of the model, which allows the decision boundary to move in the feature space. K(x,xi ) is a kernel function, which is used to calculate the similarity between user behavior characteristics and fund product attributes to determine the degree of match between them.

[0137] In an optional implementation, performing product calculation according to the first matching degree and the second matching degree to obtain a third matching degree, and using the third matching degree as a standard matching degree to recommend a fund product to a user includes:

[0138] Sorting the fund products in the third matching degree from large to small, and generating a recommendation list, and recommending the fund products in the recommendation list to the user in order from large to small;

[0139] It should be noted that an embodiment of the present invention provides an investment dynamic recommendation system based on matching analysis for executing all process steps of a recommendation method for investment in the above embodiment. The working principles and beneficial effects of the two correspond one to one, and thus will not be described in detail.

[0140] The embodiment of the present invention also provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a support vector machine classification program. When the processor executes the computer program, the steps in the above-mentioned investment recommendation method embodiment are implemented, such as Figure 1 Alternatively, when the processor executes the computer program, the functions of the modules in the above-mentioned system embodiments are realized, such as the user behavior feature matching module.

[0141] Exemplarily, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the electronic device.

[0142] The electronic device may be a computing device such as a desktop computer, a notebook, a PDA, and a smart tablet. The electronic device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. The electronic device may include more or fewer components than the above components, or a combination of certain components, or different components. For example, the electronic device may also include an input / output device, a network access device, a bus, etc.

[0143] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the electronic device, and uses various interfaces and lines to connect various parts of the entire electronic device.

[0144] The memory can be used to store the computer programs and modules, and the processor realizes various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0145] Wherein, if the module / unit integrated in the electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0146] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0147] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A recommendation method for investment, characterized in that: Executed by computer systems, including: Obtain fund product information and user behavior characteristics, and establish a fund product information database and a user behavior characteristics database; Performing a first weight calculation based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund; Perform a second weight calculation based on the matching attribute weights to obtain quantized data of user behavior characteristics and fund product characteristics; Performing a first matching degree calculation based on the feature vector quantized data to obtain a first matching degree; wherein the first matching degree is a matching degree between the fund product information database and the user behavior feature database; According to the user behavior feature database, based on support vector machine classification, a second matching degree calculation is performed to obtain a second matching degree, wherein the second matching degree is a matching degree between the user behavior feature and the fund product information; A product calculation is performed based on the first matching degree and the second matching degree to obtain a third matching degree, and the fund product is recommended to the user based on the third matching degree as the standard matching degree.

2. A method for recommending investment according to claim 1, characterized in that: The fund product information database includes project code, project name, manager code and the total number of all fund products; The user behavior feature database includes the number of all fund products browsed by the user, the code of the fund product most recently browsed by the user, the time the user browsed a single fund product, the total time the user browsed fund products, the codes of the fund products in which the user invested, the fund products purchased by the user and the amount of investment products purchased by the user.

3. A method for recommending investment according to claim 1, characterized in that: The first weight calculation is performed according to the fund product information database and the user behavior feature database to obtain the matching attribute weight between the user and the fund, including: The matching attribute weight is calculated by the following formula: W1=w×u×u2 Among them, W1 is the matching attribute weight between users and funds, w is the time weight of users browsing products, u is the number weight of funds, and u2 is the investment amount weight; Where w = T / T0, u = V / U0, u2 = a × A; Among them, T is the time the user browses a single fund product, T0 is the total time the user browses fund products, V is the number of all fund products browsed by the user, U0 is the total number of all fund products in the fund product information database, a is the preset amount ratio coefficient used to standardize the amount value, and A is the amount of the user purchasing the investment product.

4. A method for recommending investment according to claim 1, characterized in that: The second weight calculation is performed according to the matching attribute weight to obtain the user behavior characteristics and the characteristic vector quantized data of the fund product, including: A formula is calculated based on the matching attribute weight and the time weight of the user browsing the product to obtain the time feature weight of the user browsing the product; According to the time feature weight, the fund products are classified based on the support vector machine to obtain various categories of fund products, and the standard deviation of each category is calculated to obtain user behavior characteristics and feature vector quantized data of fund products.

5. A method for recommending investment according to claim 1, characterized in that: The step of calculating a first matching degree according to the feature vectorized data to obtain a first matching degree includes: The first matching degree calculation is a cosine similarity calculation, and the specific formula is: Among them, P is the matching degree between the fund product information database and the user behavior feature database, A is the feature vector quantized data of the fund product, and B is the feature vector quantized data of the user behavior feature.

6. A method for investment recommendation according to claim 1, characterized in that: The step of performing a second matching degree calculation based on the user behavior feature database and support vector machine classification to obtain a second matching degree includes: The formula for calculating the second matching degree is: Among them, Q is the matching degree between user behavior characteristics and fund product information, x is user behavior characteristics, and x i is the fund product information, n is the number of support vectors, y i is the label of the i-th support vector, α i is the Lagrange multiplier of the i-th training sample, which represents the weight of the i-th training sample in the final decision function. b is the bias term of the model, which allows the decision boundary to move in the feature space. K(x,x i ) is a kernel function, which is used to calculate the similarity between user behavior characteristics and fund product attributes to determine the degree of match between them.

7. A method for investment recommendation according to claim 1, characterized in that: The method of performing a product calculation based on the first matching degree and the second matching degree to obtain a third matching degree, and using the third matching degree as a standard matching degree to recommend a fund product to a user includes: Sorting the fund products in the third matching degree from large to small, and generating a recommendation list, and recommending the fund products in the recommendation list to the user in order from large to small; Wherein, the formula for calculating the product is: S=P×Q Among them, S is the third matching degree, Q is the second matching degree, that is, the matching degree between the user behavior characteristics and the fund product information, and P is the first matching degree, that is, the matching degree between the fund product information database and the user behavior characteristics database.

8. An investment dynamic recommendation system based on matching analysis, characterized in that: include: The data acquisition module is used to acquire fund product information and user behavior characteristics, and establish a fund product information database and a user behavior characteristics database; An attribute weight module, used to perform a first weight calculation based on the fund product information database and the user behavior feature database to obtain a matching attribute weight between the user and the fund; A feature vectorization module, used to perform a second weight calculation according to the matching attribute weight to obtain feature vectorization data of user behavior features and fund products; A fund product matching module, configured to calculate a first matching degree according to the feature vector quantized data to obtain a first matching degree, wherein the first matching degree is a matching degree between the fund product information database and the user behavior feature database; A user behavior feature matching module, configured to perform a second matching degree calculation based on the user behavior feature database and support vector machine classification to obtain a second matching degree, wherein the second matching degree is a matching degree between the user behavior feature and the fund product information; The product recommendation module is used to perform product calculation based on the first matching degree and the second matching degree to obtain a third matching degree, and recommend the fund product to the user based on the third matching degree as the standard matching degree.

9. An electronic device, characterized in that: The invention comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a recommendation method for investment as claimed in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute a recommendation method for investment as described in any one of claims 1 to 7.

Images