Fund growth evaluation method and device and computer readable storage medium

Abstract

The invention belongs to the technical field of fund evaluation, and relates to a fund growth evaluation method and device, and a computer readable storage medium. The method comprises the following steps: inputting a historical return rate data sequence of a to-be-evaluated fund into a time sequence prediction model, and outputting a predicted return rate sequence; calculating the decision-making efficiency of the to-be-assessed fund based on the historical decision-making time and the market reference decision-making time of the to-be-assessed fund; performing feature extraction on the historical evaluation data of the to-be-evaluated fund to obtain an emotion score of the to-be-evaluated fund; calculating a preference index of the to-be-assessed fund based on the emotion score and the fluctuation rate of the to-be-assessed fund; based on the investor overlapping degree of the to-be-assessed fund and each fund product, constructing a market fund network relation graph and calculating a centrality index of the to-be-assessed fund; and calculating a growth score of the to-be-evaluated fund based on the predicted return rate sequence, the decision efficiency, the preference index and the centrality index.

Description

Technical Field

[0001] The present invention relates to the technical field of fund evaluation, and in particular to a fund growth evaluation method, device and computer-readable storage medium. Background Art

[0002] With the development of financial technology and the increasing demand of investors for scientific investment methods, fund growth evaluation has gradually become an important research direction in the financial field. It can not only help investors screen fund products with high growth potential, but also help fund managers optimize fund management strategies and improve the market competitiveness of funds.

[0003] Most of the existing fund growth evaluation methods are based on the historical financial data of the fund products themselves to evaluate their potential profitability. For example, the profitability of the fund is evaluated by using the fund's rate of return in the past few years, the volatility of the fund in the market environment is measured by calculating the standard deviation of the fund's historical returns, or the weighted average of the fund's historical net asset growth rate, volatility and Sharpe ratio is used to measure the fund's growth. However, evaluating the growth of a fund based only on the fund's historical data and financial data has the following defects: 1. Historical data can only reflect the fund's performance in the past market environment, and cannot reflect the fund's growth potential in the future. Therefore, the evaluation results based on historical data cannot fully reflect the fund's future growth; 2. The existing evaluation methods only evaluate from the fund's own data, with a single evaluation angle, ignoring the decision-making data in the fund's decision-making process and the impact of the correlation between multiple funds in the market on the fund's growth evaluation, resulting in low accuracy of the evaluation results, which makes investors and fund managers have wrong expectations of the fund's potential returns and risks, and then make wrong investment decisions.

[0004] In summary, the existing fund growth evaluation method evaluates the fund's growth based on the fund's historical financial data. The evaluation results obtained cannot fully reflect the fund's future growth potential, and the evaluation angle is single, ignoring the decision-making data in the fund decision-making process and the impact of the correlation between multiple funds in the market on the fund's growth evaluation, resulting in low accuracy of the evaluation results. Summary of the invention

[0005] To this end, the technical problem to be solved by the present invention is to overcome the problem that the fund growth evaluation method in the prior art cannot fully reflect the fund's future growth potential, and the evaluation angle is single, ignoring the decision data in the fund decision-making process and the impact of the correlation between multiple funds in the market on the fund growth evaluation, resulting in low accuracy of the evaluation results.

[0006] In order to solve the above technical problems, the present invention provides a fund growth evaluation method, comprising: Input the historical yield data sequence of the fund to be evaluated into the trained time series prediction model, and output the predicted yield sequence of the fund to be evaluated; Calculate the decision efficiency of the fund to be evaluated based on the historical decision time of the fund to be evaluated and the market benchmark decision time; Input the historical evaluation data of the fund to be evaluated into the Word2Vec model for feature extraction to obtain a text feature vector; input the text feature vector into the regression support vector machine model to output the sentiment score of the fund to be evaluated; obtain the preference index of the fund to be evaluated based on the weighted sum of the sentiment score and the volatility of the fund to be evaluated; Based on the investor overlap between the fund to be evaluated and various fund products on the market, a market fund network relationship diagram is constructed, and based on the market fund network relationship diagram, a centrality index of the fund to be evaluated is calculated; Based on the predicted rate of return sequence, the decision efficiency, the preference index and the centrality index, the growth score of the fund to be evaluated is calculated.

[0007] Preferably, the loss function when training the time series prediction model is: , in, Represents the loss function for training the time series forecasting model; Represents the mean square error loss between the predicted yield series and the actual yield series; represents the predicted return series; represents the real rate of return series; represents the weight of the regularization loss term for predicting return volatility; represents the regularization loss term of the predicted return volatility; Represents the tth predicted rate of return value in the predicted rate of return sequence; Represents the t-1th predicted rate of return value in the predicted rate of return sequence; , in, Represents the number of predicted returns in the predicted return series.

[0008] Preferably, the calculation formula for the decision-making efficiency of the fund to be evaluated is: , in, It represents the decision-making efficiency of the fund to be evaluated; Indicates the number of historical decisions of the fund to be evaluated; Indicates the fund to be evaluated Historical decision time; Indicates the market benchmark decision time.

[0009] Preferably, the calculation formula of the preference index of the fund to be evaluated is: , in, represents the preference index of the fund to be evaluated; represents the sentiment score of the fund to be evaluated; express The weight of It represents the volatility of the fund to be evaluated; express The weight of .

[0010] Preferably, a market fund network relationship diagram is constructed based on the investor overlap between the fund to be evaluated and various fund products on the market, and the centrality index of the fund to be evaluated is calculated based on the market fund network relationship diagram, including: Based on the holding data of the fund to be evaluated and various fund products on the market, obtain the investor set of the fund to be evaluated and various fund products; The fund to be evaluated and each fund product on the market are taken as nodes, the Jaccard similarity between any two nodes is calculated based on the investor set of each node, and two nodes whose Jaccard similarity is greater than a preset threshold are connected to obtain a market fund network relationship diagram; Calculate the degree centrality, closeness centrality and betweenness centrality of the fund to be evaluated based on the market fund network relationship diagram; Based on the weighted sum of the degree centrality, closeness centrality and betweenness centrality of the fund to be evaluated, a centrality index of the fund to be evaluated is obtained.

[0011] Preferably, the calculation formula of the Jaccard similarity between two nodes is: , in, Representation Node and nodes Jaccard similarity; Representation Node A collection of investors; Representation Node A collection of investors; The calculation formula of the degree centrality of the fund to be evaluated is: , in, Indicates funds to be evaluated The degree centrality of Indicates the market fund network relationship diagram and the fund to be evaluated The number of connected nodes; Represents the total number of nodes in the market fund network relationship graph; The calculation formula for the closeness centrality of the fund to be evaluated is: , in, Indicates funds to be evaluated The closeness centrality of Indicates funds to be evaluated In the market fund network relationship diagram The distance between nodes; The calculation formula of the betweenness centrality of the fund to be evaluated is: , in, Indicates funds to be evaluated Betweenness centrality; Indicates the first Nodes and The nodes are connected by the fund to be evaluated The number of shortest paths; Indicates the first Nodes and The number of shortest paths between nodes; The calculation formula of the centrality index of the fund to be evaluated is: , in, represents the centrality index of the fund to be evaluated; represents the weight of the degree centrality of the fund to be evaluated; represents the weight of the proximity centrality of the fund to be evaluated; The weight representing the betweenness centrality of the fund to be evaluated.

[0012] Preferably, based on the predicted rate of return sequence, the decision efficiency, the preference index and the centrality index, calculating the growth score of the fund to be evaluated includes: The predicted yield sequence, decision efficiency, preference index and centrality index of the fund to be evaluated are standardized to obtain the standard value of the predicted yield sequence, decision efficiency, preference index and centrality index of the fund to be evaluated; Based on the standard value of the predicted yield sequence, the standard value of decision efficiency, the standard value of preference index and the standard value of centrality index of the fund to be evaluated, the entropy value of the predicted yield sequence, the entropy value of decision efficiency, the entropy value of preference index and the entropy value of centrality index of the fund to be evaluated are calculated respectively; Based on the entropy of the predicted yield sequence, decision efficiency, preference index and centrality index of the fund to be evaluated, the weight of the predicted yield sequence, decision efficiency, preference index and centrality index of the fund to be evaluated are calculated; The growth score of the fund to be evaluated is calculated based on the predicted rate of return sequence and its weight, decision efficiency and its weight, preference index and its weight, and centrality index and its weight.

[0013] Preferably, the calculation formula for the growth score of the fund to be evaluated is: , in, Indicates the growth score of the fund to be evaluated; represents the predicted return series; Represents the weight of the predicted return series; Indicates decision efficiency; The weight representing the decision efficiency; Indicates preference index; represents the weight of the preference index; represents the centrality index; represents the weight of the centrality index.

[0014] The present invention also provides a fund growth evaluation device, comprising: The yield prediction module is used to input the historical yield data sequence of the fund to be evaluated into the trained time series prediction model and output the predicted yield sequence of the fund to be evaluated; A decision efficiency calculation module, used to calculate the decision efficiency of the fund to be evaluated based on the historical decision time of the fund to be evaluated and the market benchmark decision time; The preference index calculation module is used to input the historical evaluation data of the fund to be evaluated into the Word2Vec model for feature extraction to obtain a text feature vector; input the text feature vector into the regression support vector machine model to output the sentiment score of the fund to be evaluated; based on the weighted sum of the sentiment score and the volatility of the fund to be evaluated, obtain the preference index of the fund to be evaluated; A centrality index calculation module is used to construct a market fund network relationship diagram based on the investor overlap between the fund to be evaluated and various fund products on the market, and calculate the centrality index of the fund to be evaluated based on the market fund network relationship diagram; The growth score evaluation module is used to calculate the growth score of the fund to be evaluated based on the predicted rate of return sequence, the decision efficiency, the preference index and the centrality index.

[0015] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned fund growth evaluation method are implemented.

[0016] The fund growth evaluation method provided in this application first predicts the future rate of return based on the historical rate of return of the fund to be evaluated. Since the decision-making efficiency of the fund can reflect the decision-making difficulty of the fund to be evaluated to a certain extent, and then reflect the market adaptability of the fund to be evaluated, this application calculates the decision-making efficiency based on the decision-making time and market benchmark time of the fund to be evaluated. At the same time, the evaluation data of the fund to be evaluated can reflect the performance stability and development potential of the fund to be evaluated. Therefore, this application extracts the sentiment score of the fund to be evaluated from the historical evaluation data, combines it with the volatility of the fund to be evaluated to obtain the preference index of the fund to be evaluated, and reflects the development potential of the fund to be evaluated by the level of the preference index; finally, considering various There will be direct or indirect connections between fund products. The more funds a fund is associated with, the more important the fund is in the market and the higher its growth potential is. Therefore, this application constructs a market fund network relationship diagram based on the investor overlap of each fund, and reflects the importance of the fund to be evaluated in the market by calculating the centrality index of the fund to be evaluated; this application calculates the growth score of the fund based on the predicted rate of return, decision-making efficiency and preference index of the fund to be evaluated and the market importance of the fund to be evaluated, and evaluates the growth score of the fund more comprehensively and accurately, which can enable investors and fund managers to more accurately anticipate the fund's future investment returns and risks, and thus make more accurate investment choices. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below according to specific embodiments of the present invention in conjunction with the accompanying drawings, wherein: Figure 1 A flow chart of the fund growth assessment method provided for this application; Figure 2 This is a schematic diagram of the structure of the fund growth assessment device provided in this application. DETAILED DESCRIPTION

[0018] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0019] See also Figure 1 , Figure 1 The following is a flow chart of the fund growth assessment method provided in this application, which specifically includes: S10: Input the historical yield data sequence of the fund to be evaluated into the trained time series prediction model, and output the predicted yield sequence of the fund to be evaluated; S20: Calculate the decision efficiency of the fund to be evaluated based on the historical decision time of the fund to be evaluated and the market benchmark decision time; S30: Input the historical evaluation data of the fund to be evaluated into the Word2Vec model for feature extraction to obtain a text feature vector; input the text feature vector into the regression support vector machine model to output the sentiment score of the fund to be evaluated; obtain the preference index of the fund to be evaluated based on the weighted sum of the sentiment score and the volatility of the fund to be evaluated; S40: Based on the investor overlap between the fund to be evaluated and various fund products on the market, a market fund network relationship diagram is constructed, and the centrality index of the fund to be evaluated is calculated based on the market fund network relationship diagram; S50: Calculate the growth score of the fund to be evaluated based on the predicted yield sequence, decision efficiency, preference index and centrality index.

[0020] The fund growth evaluation method provided in this application first predicts the future rate of return based on the historical rate of return of the fund to be evaluated. Since the decision-making efficiency of the fund can reflect the decision-making difficulty of the fund to be evaluated to a certain extent, and then reflect the market adaptability of the fund to be evaluated, this application calculates the decision-making efficiency based on the decision-making time and market benchmark time of the fund to be evaluated. At the same time, the evaluation data of the fund to be evaluated can reflect the performance stability and development potential of the fund to be evaluated. Therefore, this application extracts the sentiment score of the fund to be evaluated from the historical evaluation data, combines it with the volatility of the fund to be evaluated to obtain the preference index of the fund to be evaluated, and reflects the development potential of the fund to be evaluated by the level of the preference index; finally, considering various There will be direct or indirect connections between fund products. The more funds a fund is associated with, the more important the fund is in the market and the higher its growth potential is. Therefore, this application constructs a market fund network relationship diagram based on the investor overlap of each fund, and reflects the importance of the fund to be evaluated in the market by calculating the centrality index of the fund to be evaluated; this application calculates the growth score of the fund based on the predicted rate of return, decision-making efficiency and preference index of the fund to be evaluated and the market importance of the fund to be evaluated, and evaluates the growth score of the fund more comprehensively and accurately, which can enable investors and fund managers to more accurately anticipate the fund's future investment returns and risks, and thus make more accurate investment choices.

[0021] Since the fund's rate of return is an important data for evaluating fund performance, it reflects the fund's return level and helps investors evaluate the fund's potential profitability. Investors can choose funds that meet their own risks based on the fund's expected return. Therefore, this application uses the fund's rate of return as a reference indicator when evaluating the fund's growth. However, since the fund's historical rate of return can only reflect the fund's past returns and cannot reflect the fund's future growth potential, this application predicts the fund's future rate of return based on historical rate of return data and uses the predicted rate of return as an evaluation indicator to reflect the fund's growth potential and return level to a certain extent.

[0022] Specifically, the historical yield data sequence of the fund to be evaluated can be obtained on the financial data platform, and the historical yield data sequence can be input into the trained time series prediction model to output the predicted yield sequence of the fund to be evaluated in the future.

[0023] Optionally, the time series prediction model can be a linear regression model, an LSTM model, a random forest model, etc.

[0024] In some embodiments of the present application, the loss function when training the time series prediction model is: , in, Represents the loss function for training the time series forecasting model; Represents the mean square error loss between the predicted yield series and the actual yield series; represents the predicted return series; represents the real rate of return series; represents the weight of the regularization loss term for predicting return volatility; represents the regularization loss term of the predicted return volatility; Represents the tth predicted rate of return value in the predicted rate of return sequence; Represents the t-1th predicted rate of return value in the predicted rate of return sequence; , in, Represents the number of predicted returns in the predicted return series.

[0025] During the yield prediction process, the model will pay too much attention to the abnormally high returns or extreme fluctuations in the historical yield data. If the volatility of the historical yield data is ignored, the model will overfit the high yield intervals in the historical yield data, resulting in inaccurate predicted yield data. Based on this, in addition to paying attention to the mean square error between the predicted yield and the actual yield in the loss function, this application also adds a predicted yield volatility regularization loss term to ensure that the yield curve can be smoothed during the prediction process, avoiding large fluctuations in the predicted yield sequence caused by overfitting of the model, and making the predicted yield sequence more reasonable.

[0026] Furthermore, the historical decision-making time of the fund to be evaluated refers to the time required for the fund to be evaluated to make an investment decision after a key market event (such as the release of economic data, policy changes, market crash, etc.); the shorter the historical decision-making time, the easier it is for the fund to be evaluated to make a decision, which to a certain extent reflects that the fund to be evaluated has good market adaptability and will be easier to make decisions when encountering similar market changes in the future. The longer the historical decision-making time, the poorer the market adaptability of the fund to be evaluated. Therefore, this application compares the historical decision-making time of the fund to be evaluated with the market benchmark decision time, and then obtains the decision-making efficiency of the fund to be evaluated, so as to reflect the market adaptability of the fund to be evaluated.

[0027] Specifically, the calculation formula for the decision-making efficiency of the fund to be evaluated is: , in, It represents the decision-making efficiency of the fund to be evaluated; Indicates the number of historical decisions of the fund to be evaluated; Indicates the fund to be evaluated Historical decision time; Indicates the market benchmark decision time.

[0028] In step S30, the sentiment score of the fund to be evaluated is obtained by using the Word2Vec model and the regression support vector machine model, and then the preference index of the fund to be evaluated is calculated based on the sentiment score and the volatility of the fund to be evaluated. Since the sentiment score of the fund to be evaluated can also reflect the performance stability of the fund to be evaluated to a certain extent, if the sentiment score of the fund to be evaluated is high, it indicates that the performance of the fund to be evaluated may be better than the market benchmark, and in the long run it can provide continuous positive returns and maintain relatively stable growth; if the sentiment score of the fund to be evaluated is low, it indicates that the performance stability of the fund to be evaluated is poor and the return is low.

[0029] Specifically, the calculation formula for the preference index of the fund to be evaluated is: , in, represents the preference index of the fund to be evaluated; represents the sentiment score of the fund to be evaluated; express The weight of It represents the volatility of the fund to be evaluated; express The weight of .

[0030] Furthermore, in addition to predicting the future rate of return data of the fund to be evaluated and the evaluation index of the fund to be evaluated, the present application also reflects the growth of the fund to be evaluated by calculating the importance of the fund to be evaluated in the fund market. If the centrality index of the fund to be evaluated in the market fund network is high, it indicates that the fund to be evaluated has a direct connection with other funds, which can further reflect that the fund to be evaluated is an important fund in the market. Specifically, the implementation method of step S40 includes: S401: Based on the holding data of the fund to be evaluated and various fund products on the market, obtain the investor set of the fund to be evaluated and various fund products; S402: The fund to be evaluated and each fund product in the market are taken as nodes, the Jaccard similarity between any two nodes is calculated based on the investor set of each node, and two nodes whose Jaccard similarity is greater than a preset threshold are connected to obtain a market fund network relationship diagram; Specifically, the calculation formula of the Jaccard similarity between two nodes is: , in, Representation Node and nodes Jaccard similarity; Representation Node A collection of investors; Representation Node A collection of investors; S403: Calculating the degree centrality, closeness centrality and betweenness centrality of the fund to be evaluated based on the market fund network relationship diagram; Specifically, the calculation formula for the degree centrality of the fund to be evaluated is: , in, Indicates funds to be evaluated The degree centrality of Indicates the market fund network relationship diagram and the fund to be evaluated The number of connected nodes; Represents the total number of nodes in the market fund network relationship graph; The calculation formula for the closeness centrality of the fund to be evaluated is: , in, Indicates funds to be evaluated The closeness centrality of Indicates funds to be evaluated In the market fund network relationship diagram The distance between nodes; The calculation formula of the betweenness centrality of the fund to be evaluated is: , in, Indicates funds to be evaluated Betweenness centrality; Indicates the first Nodes and The nodes are connected by the fund to be evaluated The number of shortest paths; Indicates the first Nodes and The number of shortest paths between nodes; S404: obtaining a centrality index of the fund to be evaluated based on a weighted sum of the degree centrality, closeness centrality and betweenness centrality of the fund to be evaluated; Specifically, the calculation formula for the centrality index of the fund to be evaluated is: , in, represents the centrality index of the fund to be evaluated; represents the weight of the degree centrality of the fund to be evaluated; represents the weight of the proximity centrality of the fund to be evaluated; The weight representing the betweenness centrality of the fund to be evaluated.

[0031] Optionally, in some embodiments of the present application, the investment weight of each investor in the investor set of each fund product can also be obtained, and the Pearson correlation coefficient of the two funds is calculated based on the investment weight of each investor in the investor set of the two funds. The Pearson correlation coefficient is used to characterize the asset overlap of the two funds, thereby constructing a market fund network relationship diagram.

[0032] Furthermore, in order to avoid artificially setting the subjective weight of each indicator, which may lead to a lack of objectivity in the scoring result and affect the accuracy of the evaluation result, this application adopts the entropy weight method to determine the objective weight of each indicator when calculating the growth score of the fund to be evaluated based on each indicator. Specifically, the implementation process of step S50 includes: S501: performing standardization processing on the predicted rate of return sequence, decision efficiency, preference index and centrality index of the fund to be evaluated, and obtaining the standard value of the predicted rate of return sequence, the standard value of decision efficiency, the standard value of preference index and the standard value of centrality index of the fund to be evaluated; S502: based on the standard value of the predicted yield sequence, the standard value of decision efficiency, the standard value of preference index and the standard value of centrality index of the fund to be evaluated, respectively calculating the entropy value of the predicted yield sequence, the entropy value of decision efficiency, the entropy value of preference index and the entropy value of centrality index of the fund to be evaluated; Specifically, the calculation formula of the entropy value of each indicator is: , in, Indicates The entropy value of each indicator; Represents the number of nodes in the market fund network relationship graph; Indicates The indicator is in the market fund network relationship diagram The standardized value in the fund corresponding to each node; S503: Calculate the weight of the predicted yield sequence, the weight of the decision efficiency, the weight of the preference index and the weight of the centrality index of the fund to be evaluated based on the entropy value of the predicted yield sequence, the entropy value of the decision efficiency, the entropy value of the preference index and the entropy value of the centrality index of the fund to be evaluated; Specifically, the calculation formula for the weight of each indicator is: , in, Indicates The weight of each indicator; Indicates the number of indicators; S504: Calculate the growth score of the fund to be evaluated based on the predicted rate of return sequence and its weight, decision efficiency and its weight, preference index and its weight, and centrality index and its weight.

[0033] Specifically, the calculation formula for the growth score of the fund to be evaluated is: , in, Indicates the growth score of the fund to be evaluated; represents the predicted return series; Represents the weight of the predicted return series; Indicates decision efficiency; The weight representing the decision efficiency; Indicates preference index; represents the weight of the preference index; represents the centrality index; represents the weight of the centrality index.

[0034] Based on the fund growth evaluation method provided in the above embodiment, the present application embodiment also provides a fund growth evaluation device, such as Figure 2 As shown, the device specifically includes: The yield prediction module 10 is used to input the historical yield data sequence of the fund to be evaluated into the trained time series prediction model and output the predicted yield sequence of the fund to be evaluated; A decision efficiency calculation module 20, for calculating the decision efficiency of the fund to be evaluated based on the historical decision time of the fund to be evaluated and the market benchmark decision time; The preference index calculation module 30 is used to input the historical evaluation data of the fund to be evaluated into the Word2Vec model for feature extraction to obtain a text feature vector; input the text feature vector into the regression support vector machine model to output the sentiment score of the fund to be evaluated; based on the weighted sum of the sentiment score and the volatility of the fund to be evaluated, obtain the preference index of the fund to be evaluated; The centrality index calculation module 40 is used to construct a market fund network relationship diagram based on the investor overlap between the fund to be evaluated and various fund products on the market, and calculate the centrality index of the fund to be evaluated based on the market fund network relationship diagram; The growth score evaluation module 50 is used to calculate the growth score of the fund to be evaluated based on the predicted rate of return sequence, decision efficiency, preference index and centrality index.

[0035] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned fund growth evaluation method are implemented.

[0036] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0037] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0038] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0039] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0040] Obviously, the above embodiments are merely examples for the purpose of clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the present invention.

Claims

1. A fund growth evaluation method, characterized in that: include: Input the historical yield data sequence of the fund to be evaluated into the trained time series prediction model, and output the predicted yield sequence of the fund to be evaluated; Calculate the decision efficiency of the fund to be evaluated based on the historical decision time of the fund to be evaluated and the market benchmark decision time; Input the historical evaluation data of the fund to be evaluated into the Word2Vec model for feature extraction to obtain a text feature vector; input the text feature vector into the regression support vector machine model to output the sentiment score of the fund to be evaluated; Obtaining a preference index of the fund to be evaluated based on a weighted sum of the sentiment score and the volatility of the fund to be evaluated; Based on the investor overlap between the fund to be evaluated and various fund products on the market, a market fund network relationship diagram is constructed, and based on the market fund network relationship diagram, a centrality index of the fund to be evaluated is calculated; Based on the predicted rate of return sequence, the decision efficiency, the preference index and the centrality index, the growth score of the fund to be evaluated is calculated.

2. The fund growth evaluation method according to claim 1, characterized in that: The loss function when training the time series prediction model is: ， in, Represents the loss function for training the time series forecasting model; Represents the mean square error loss between the predicted yield series and the actual yield series; represents the predicted return series; represents the real rate of return series; represents the weight of the regularization loss term for predicting return volatility; represents the regularization loss term of the predicted return volatility; Represents the tth predicted rate of return value in the predicted rate of return sequence; Represents the t-1th predicted rate of return value in the predicted rate of return sequence; ， in, Represents the number of predicted returns in the predicted return series.

3. The fund growth evaluation method according to claim 1, characterized in that: The calculation formula for the decision-making efficiency of the fund to be evaluated is: ， in, It represents the decision-making efficiency of the fund to be evaluated; Indicates the number of historical decisions of the fund to be evaluated; Indicates the fund to be evaluated Historical decision time; Indicates the market benchmark decision time.

4. The fund growth evaluation method according to claim 1, characterized in that: The calculation formula of the preference index of the fund to be evaluated is: ， in, represents the preference index of the fund to be evaluated; represents the sentiment score of the fund to be evaluated; express The weight of It represents the volatility of the fund to be evaluated; express The weight of .

5. The fund growth evaluation method according to claim 1, characterized in that: A market fund network relationship diagram is constructed based on the investor overlap between the fund to be evaluated and various fund products on the market. The centrality index of the fund to be evaluated is calculated based on the market fund network relationship diagram, including: Based on the holding data of the fund to be evaluated and various fund products on the market, obtain the investor set of the fund to be evaluated and various fund products; The fund to be evaluated and each fund product on the market are taken as nodes, the Jaccard similarity between any two nodes is calculated based on the investor set of each node, and two nodes whose Jaccard similarity is greater than a preset threshold are connected to obtain a market fund network relationship diagram; Calculate the degree centrality, closeness centrality and betweenness centrality of the fund to be evaluated based on the market fund network relationship diagram; Based on the weighted sum of the degree centrality, closeness centrality and betweenness centrality of the fund to be evaluated, a centrality index of the fund to be evaluated is obtained.

6. The fund growth evaluation method according to claim 5, characterized in that: The calculation formula for the Jaccard similarity between two nodes is: ， in, Representation Node and nodes Jaccard similarity; Representation Node A collection of investors; Representation Node A collection of investors; The calculation formula of the degree centrality of the fund to be evaluated is: ， in, Indicates funds to be evaluated The degree centrality of Indicates the market fund network relationship diagram and the fund to be evaluated The number of connected nodes; Represents the total number of nodes in the market fund network relationship graph; The calculation formula for the closeness centrality of the fund to be evaluated is: ， in, Indicates funds to be evaluated The closeness centrality of Indicates funds to be evaluated In the market fund network relationship diagram The distance between nodes; The calculation formula of the betweenness centrality of the fund to be evaluated is: ， in, Indicates funds to be evaluated Betweenness centrality; Indicates the first Nodes and The nodes are connected by the fund to be evaluated The number of shortest paths; Indicates the first Nodes and The number of shortest paths between nodes; The calculation formula of the centrality index of the fund to be evaluated is: ， in, represents the centrality index of the fund to be evaluated; represents the weight of the degree centrality of the fund to be evaluated; represents the weight of the proximity centrality of the fund to be evaluated; The weight representing the betweenness centrality of the fund to be evaluated.

7. The fund growth evaluation method according to claim 1, characterized in that: Calculating the growth score of the fund to be evaluated based on the predicted rate of return sequence, the decision efficiency, the preference index and the centrality index includes: The predicted yield sequence, decision efficiency, preference index and centrality index of the fund to be evaluated are standardized to obtain the standard value of the predicted yield sequence, decision efficiency, preference index and centrality index of the fund to be evaluated; Based on the standard value of the predicted yield sequence, the standard value of decision efficiency, the standard value of preference index and the standard value of centrality index of the fund to be evaluated, the entropy value of the predicted yield sequence, the entropy value of decision efficiency, the entropy value of preference index and the entropy value of centrality index of the fund to be evaluated are calculated respectively; Based on the entropy of the predicted yield sequence, decision efficiency, preference index and centrality index of the fund to be evaluated, the weight of the predicted yield sequence, decision efficiency, preference index and centrality index of the fund to be evaluated are calculated; The growth score of the fund to be evaluated is calculated based on the predicted rate of return sequence and its weight, decision efficiency and its weight, preference index and its weight, and centrality index and its weight.

8. The fund growth evaluation method according to claim 1, characterized in that: The calculation formula for the growth score of the fund to be evaluated is: ， in, Indicates the growth score of the fund to be evaluated; represents the predicted return series; Represents the weight of the predicted return series; Indicates decision efficiency; The weight representing the decision efficiency; Indicates preference index; represents the weight of the preference index; represents the centrality index; represents the weight of the centrality index.

9. A fund growth evaluation device, characterized in that: include: The yield prediction module is used to input the historical yield data sequence of the fund to be evaluated into the trained time series prediction model and output the predicted yield sequence of the fund to be evaluated; A decision efficiency calculation module, used to calculate the decision efficiency of the fund to be evaluated based on the historical decision time of the fund to be evaluated and the market benchmark decision time; The preference index calculation module is used to input the historical evaluation data of the fund to be evaluated into the Word2Vec model for feature extraction to obtain a text feature vector; input the text feature vector into the regression support vector machine model to output the sentiment score of the fund to be evaluated; based on the weighted sum of the sentiment score and the volatility of the fund to be evaluated, obtain the preference index of the fund to be evaluated; A centrality index calculation module is used to construct a market fund network relationship diagram based on the investor overlap between the fund to be evaluated and various fund products on the market, and calculate the centrality index of the fund to be evaluated based on the market fund network relationship diagram; The growth score evaluation module is used to calculate the growth score of the fund to be evaluated based on the predicted rate of return sequence, the decision efficiency, the preference index and the centrality index.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the fund growth evaluation method according to any one of claims 1 to 8 is implemented.

Images