Intelligent recommendation method and device, equipment, storage medium and program product

Abstract

The invention discloses an intelligent recommendation method and device, equipment, a storage medium and a program product, and the method comprises the steps: determining a first similarity matrix between a plurality of ETFs and a plurality of indexes based on the constituent stock information of the ETFs and the constituent stock information of the indexes; based on target stock information and the first similarity matrix, determining a plurality of second similarity results between the target stock and the ETFs; and on the basis of the plurality of second similarity results, determining one or more target ETFs, and recommending the one or more target ETFs.

Description

Technical Field

[0001] This application relates to the fields of machine learning and financial data services, and particularly to an intelligent recommendation method, apparatus, device, storage medium, and program product. Background Art

[0002] Exchange Traded Funds (ETFs) have become one of the important choices for more and more investors' asset allocation due to their advantages such as being able to track specific indices, relatively low transaction costs, and flexible trading. Currently, the original design of ETFs is to closely track various indices, and their constituent stocks are usually selected according to the compilation rules of the corresponding indices. However, there are still some limitations in the ETF recommendation methods in related technologies, resulting in insufficient tightness of the correlation between the recommended ETFs and stocks, and being unable to accurately meet the needs of investors for ETFs with highly similar investment characteristics to the stocks they are concerned about. Summary of the Invention

[0003] To solve the above technical problems, embodiments of the present invention provide an intelligent recommendation method, apparatus, device, storage medium, and program product.

[0004] The intelligent recommendation method provided by an embodiment of this application, which is applied to a big data warehouse, includes:

[0005] Based on the constituent stock information of multiple Exchange Traded Funds (ETFs) and the constituent stock information of multiple indices, determining a first similarity matrix between the multiple ETFs and the multiple indices;

[0006] Based on the target stock information and the first similarity matrix, determining multiple second similarity results between the target stock and the multiple ETFs;

[0007] Based on the multiple second similarity results, determining one or more target ETFs and recommending the one or more target ETFs.

[0008] The intelligent recommendation apparatus provided by an embodiment of this application, which is applied to a big data warehouse, includes:

[0009] A first determination unit, configured to determine a first similarity matrix between the multiple ETFs and the multiple indices based on the constituent stock information of multiple Exchange Traded Funds (ETFs) and the constituent stock information of multiple indices;

[0010] A second determination unit, configured to determine multiple second similarity results between the target stock and the multiple ETFs based on the target stock information and the first similarity matrix;

[0011] A third determination unit, configured to determine one or more target ETFs based on the multiple second similarity results;

[0012] A recommendation unit, configured to recommend the one or more target ETFs.

[0013] The processing device provided by the embodiment of the present application includes: a processor and a memory. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute any one of the above intelligent recommendation methods.

[0014] The computer-readable storage medium provided by the embodiment of the present application is used to store a computer program, and the computer program enables a computer to execute any one of the above intelligent recommendation methods.

[0015] The computer program product provided by the embodiment of the present application includes computer program instructions, and the computer program instructions enable a computer to execute any one of the above intelligent recommendation methods.

[0016] In the technical solution of the embodiment of the present application, based on the constituent stock information of multiple exchange-traded funds (ETFs) and the constituent stock information of multiple indexes, a first similarity matrix between multiple ETFs and multiple indexes is determined. Based on the target stock information and the first similarity matrix, multiple second similarity results between the target stock and multiple ETFs are determined. Based on the multiple second similarity results, one or more target ETFs are determined, and the one or more target ETFs are recommended. In this way, by analyzing the similarity of the constituent stocks of ETF funds and various indexes and integrating it into the evaluation system of the overall similarity between the target stock and ETFs, it can be ensured that the recommended ETFs have a strong correlation with the target stock in terms of constituent stocks, so as to be able to accurately recommend ETF funds that highly match the investment characteristics of the target stock for the target stock, meet the needs of investors for ETFs with highly similar investment characteristics to the stocks they are concerned about, and improve the accuracy and effectiveness of ETF recommendation. Description of the Drawings

[0017] Figure 1 is a schematic flowchart of an intelligent recommendation method applied to a big data warehouse provided by the embodiment of the present application;

[0018] Figure 2 is a schematic flowchart of an ETF intelligent recommendation method for stocks provided by the embodiment of the present application;

[0019] Figure 3 is a schematic structural diagram of an intelligent recommendation device applied to a big data warehouse provided by the embodiment of the present application;

[0020] Figure 4 is a schematic structural diagram of a processing device provided by the embodiment of the present application. Detailed implementation manners

[0021] The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the protection scope of the present application.

[0022] In the description of the embodiments of the present application, the term "corresponding" may indicate a direct or indirect corresponding relationship between two entities, or may indicate an associated relationship between them, or may be a relationship such as indication and being indicated, configuration and being configured, etc.

[0023] To facilitate the understanding of the technical solutions in the embodiments of the present application, the related technologies in the embodiments of the present application will be described below. The following related technologies can be arbitrarily combined with the technical solutions in the embodiments of the present application as optional solutions, and they all fall within the protection scope of the embodiments of the present application.

[0024] With the continuous development and continuous expansion of the financial market, both the stock market and the ETF market have shown an increasingly complex and diversified trend. Many investors attempt to obtain returns and diversify risks by reasonably allocating assets. In the field of stock investment, investors often need to analyze and screen a large number of stocks, and make investment decisions based on various fundamental data (such as the company's financial status, industry prospects, etc.) and technical indicators (such as moving averages, trading volumes, etc.). For ETFs, with their advantages of being able to track specific indices, relatively low trading costs, and flexible trading, they have become one of the important choices for more and more investors' asset allocation.

[0025] An ETF is an open-ended fund listed and traded on an exchange with variable fund shares, and its purpose is to provide investors with an index investment tool that can be traded in the secondary market like stocks. Since the launch of ETFs, they have developed rapidly globally and have become a popular investment variety. Currently, the original design intention of ETFs is to closely track various indices, and their constituent stocks are usually selected according to the compilation rules of the corresponding indices. Each index has a clear component ratio setting to reflect the weight of different stocks in the market segment or the overall market represented by the index.

[0026] An index is a statistical indicator used to measure and reflect the overall performance of a specific market or asset class. It comprehensively calculates the prices of a representative group of stocks, bonds, commodities, or other assets, presenting the overall trend and volatility of that market or asset class in a concise and clear manner. For example, for the common CSI 300 Index, 300 constituent stocks and their respective weights are determined according to certain screening and weighting methods. The constituent stocks of an ETF that tracks the CSI 300 Index are basically highly coincident with those of the index, except that there may be differences in minor weight adjustments and replication strategies.

[0027] In the relevant financial investment service ecosystem, there is a representative method for recommending ETFs for stocks based on industry classification. This method specifically includes the following processes:

[0028] First, the system collects a vast amount of financial data information from multiple authoritative financial data sources (such as professional financial data service companies, official databases of stock exchanges, etc.). Among them, for stock data, it covers the basic information of stocks, mainly including stock codes, stock names, listing exchanges, industry classifications (according to industry classification standards, such as the common Shenwan Level 1 industry classification, which clearly classifies stocks into different industry segments such as finance, medicine, consumption, technology, etc.). For ETF data, it mainly includes the fund codes of ETFs, fund names, the names and codes of the indexes they track, information on the management companies of ETFs, net asset values, listing and trading times, and average daily trading volumes over a certain period in the past. The key is to clarify the specific indexes tracked by each ETF and their basic trading activity and scale conditions.

[0029] Second, when an investor browses the details page of a certain stock on a relevant financial platform, the system background will extract the industry information of the stock. Then, based on this industry classification, it will screen among the collected and sorted ETF data. Specifically, it mainly searches for those ETFs whose tracked indexes clearly correspond to the industry of the stock. For example, if an investor views a stock in the pharmaceutical industry, the system will screen out ETFs that track pharmaceutical industry indexes (such as the CSI Pharmaceutical Index, Guozheng Pharmaceutical Index, and other related pharmaceutical-themed indexes) from numerous ETFs.

[0030] Then, after screening out the ETFs corresponding to the same industry, some relatively conventional sorting rules are usually adopted to determine the order of the final ETF list recommended to investors. Common sorting methods include the following:

[0031] (1) Sort by the asset size of the ETF. Larger asset size of the ETF often implies better liquidity and market recognition. Therefore, generally, ETFs with a higher asset size ranking are recommended first. For example, for the selected ETFs in the pharmaceutical industry, sort them in descending order according to their latest announced net asset value, and display the ETFs with larger sizes in front for investors to view first.

[0032] (2) Sort by the average daily trading volume of the ETF. The higher the average daily trading volume of the ETF, the higher the trading activity of the ETF in the market. It is easier for investors to complete transactions during buying and selling, and they will not face too much liquidity risk. Therefore, some platforms will sort and recommend ETFs in descending order according to their average daily trading volume.

[0033] However, the above method of recommending ETFs for stocks based on industry classification still has the following problems:

[0034] (1) Insufficient recommendation accuracy. Only screening and recommending ETFs based on industry classification without considering key factors such as the component proportion of stocks in each index and the similarity between the ETF fund and the constituent stocks of each index will result in a weak correlation between the recommended ETFs and the stocks, and cannot accurately meet the needs of investors for ETFs with highly similar investment characteristics to the stocks they are concerned about. For example, in the pharmaceutical industry, different pharmaceutical stocks have differences in sub - sectors, business models, market performances, etc. The ETFs selected simply based on the industry index may not accurately reflect these differences, resulting in a poor match between the recommended ETFs and the stocks, affecting investors' investment decisions and return expectations.

[0035] (2) Failure to consider the characteristics of constituent stocks. Not deeply exploring the similarity between the ETF fund and the constituent stocks of the corresponding index may lead to a weak correlation between the actual constituent stocks of the recommended ETFs and the target stocks although they track the same industry index. For example, some pharmaceutical industry ETFs may mainly invest in large pharmaceutical companies, while the stock that the investor is concerned about is a small and medium - sized pharmaceutical company focusing on innovative drug research and development. According to the above - mentioned recommendation method, there will be significant differences in business models, risk characteristics, etc. between the recommended ETFs and the target stocks, which cannot provide effective investment reference for investors, thus reducing the effectiveness and practicality of the recommendation.

[0036] To solve the above - mentioned technical problems and facilitate the understanding of the technical solutions of the embodiments of the present application, the technical solutions of the present application are described in detail through specific embodiments below. The above - mentioned related technologies can be combined arbitrarily with the technical solutions of the embodiments of the present application, and they all fall within the protection scope of the embodiments of the present application. The embodiments of the present application include at least some of the following contents.

[0037] An intelligent recommendation method applied to a big data warehouse is proposed in an embodiment of the present application. Figure 1 It is a schematic flowchart of the intelligent recommendation method applied to the big data warehouse provided by the embodiment of the present application. As Figure 1 shown, the method includes the following steps:

[0038] Step 101: Based on the constituent stock information of multiple ETFs and the constituent stock information of multiple indexes, determine a first similarity matrix between the multiple ETFs and the multiple indexes.

[0039] In the embodiment of the present application, in order to obtain the constituent stock similarity between multiple ETFs and multiple indexes, the similarity calculation can be performed on the constituent stock information of multiple ETFs and the constituent stock information of multiple indexes to obtain multiple constituent stock similarity results between the multiple ETFs and the multiple indexes, and a first similarity matrix between the multiple ETFs and the multiple indexes can be constructed according to the multiple constituent stock similarity results. Among them, the constituent stock similarity result can determine the coincidence degree of the constituent stocks between the ETF and the index. If there is a high coincidence degree of the constituent stocks between an ETF and an index, it can be determined that the correlation between the ETF and the index is very high, that is, the constituent stock similarity between the ETF and the index is very high.

[0040] Here, before calculating the similarity between the constituent stock information of multiple ETFs and the constituent stock information of multiple indexes, it is also necessary to obtain relevant information data first. Specifically, by establishing a data connection channel with an external information data source, and through the corresponding data interface and synchronization tool, the relevant information data is synchronized to the big data warehouse according to specific rules and frequencies, and the relevant information data is extracted in the big data warehouse to obtain target stock information, the constituent stock information of multiple indexes, and the constituent stock information of multiple ETFs.

[0041] Among them, the target stock information includes the code and name of the target stock, which can uniquely identify the target stock. The constituent stock information of multiple indexes includes the index code and constituent stock code of each index. The index code of each index is used to identify the index, and the constituent stock code of each index is used to determine which stocks constitute the index and can determine the proportion of each constituent stock in the index. The constituent stock information of multiple ETFs includes the ETF code and constituent stock code of each ETF. The ETF code of each ETF is used to identify the ETF, and the constituent stock code of each ETF is used to determine which stocks constitute the ETF and can determine the proportion of each constituent stock in the ETF.

[0042] In some embodiments, the constituent stock information of multiple ETFs includes the ETF code and constituent stock code of each ETF, and the constituent stock information of multiple indexes includes the index code and constituent stock code of each index; among them, for step 101, it may specifically include:

[0043] Based on the ETF code and constituent stock codes of each ETF, determine the proportion of the constituent stock weights of each ETF;

[0044] Based on the index code and constituent stock codes of each index, determine the proportion of the constituent stock weights of each index;

[0045] Based on the proportion of the constituent stock weights of each ETF and the proportion of the constituent stock weights of each index, determine the first similarity matrix.

[0046] Here, since the constituent stock codes of each ETF are used to determine which stocks constitute the corresponding ETF, it is possible to first determine each ETF identified based on the ETF code, and determine one or more stocks that constitute each ETF based on the constituent stock codes of each ETF, and determine the weight proportion of each stock in each ETF based on the market value of each stock in the ETF among the one or more stocks, so as to obtain the proportion of the constituent stock weights of each ETF.

[0047] Similarly, since the constituent stock codes of each index are used to determine which stocks constitute the corresponding index, it is possible to first determine each index identified based on the index code, and determine one or more stocks that constitute each index based on the constituent stock codes of each index, and determine the weight proportion of each stock in each index based on the market value of each stock in the index among the one or more stocks, so as to obtain the proportion of the constituent stock weights of each index.

[0048] After obtaining the proportion of the constituent stock weights of each ETF and the proportion of the constituent stock weights of each index, calculate the similarity of the constituent stocks between each ETF and each index based on the proportion of the constituent stock weights of each ETF and the proportion of the constituent stock weights of each index, so as to obtain multiple constituent stock similarity results between multiple ETFs and multiple indexes. Based on the multiple constituent stock similarity results, the corresponding first similarity matrix can be constructed, and this matrix is the index-ETF similarity matrix.

[0049] It should be noted that when calculating the proportion of constituent stocks of each index and each ETF, the latest data needs to be used. For example, select the index data and ETF data of the most recent quarter, so as to ensure that the weight proportion of each stock reflected in the index or ETF conforms to the actual situation of the current market. Among them, since the position adjustment of ETFs usually has a certain cycle, selecting the ETF data of a quarter can not only reflect its relatively stable constituent stock composition, but also better reflect the recent change trend.

[0050] In some embodiments, for "Based on the index code and constituent stock codes of each index, determine the proportion of the constituent stock weights of each index", it may specifically include:

[0051] Based on the index codes and constituent stock codes of each index, determine one or more stocks included in each index;

[0052] Based on each stock among the one or more stocks, calculate the weight percentage of each stock in each index to obtain the constituent stock weight percentages of each index.

[0053] Here, first, determine each index identified based on the index code, and determine one or more stocks constituting each index based on the constituent stock codes of each index. Secondly, obtain the real-time closing price of each stock among the one or more stocks through a market software, and multiply the total number of outstanding shares of each stock by its real-time closing price to obtain the market value of each stock in the index. Based on the market value of each stock in the index, the total market value of each index can be determined. Then, based on the market value of each stock in the index and the total market value of each index, determine the weight percentage of each stock in each index, so as to determine the constituent stock weight percentages of each index based on the weight percentage of each stock in each index.

[0054] For example, if stocks A, B, C, and D jointly constitute an index, the real-time closing price of stock A is 50 yuan, and the total number of its outstanding shares is 100 million, then the market value of stock A is 500 million yuan. Similarly, the market value of stock B can be calculated as 600 million yuan, the market value of stock C as 300 million yuan, and the market value of stock D as 200 million yuan according to the real-time closing prices and the total number of outstanding shares of stocks B, C, and D respectively. Then, sum up the market values of stocks A, B, C, and D to calculate that the total market value of this index is 1.6 billion yuan. Thus, based on the market values of stocks A, B, C, D and the total market value of this index, the weight percentages of stocks A, B, C, and D in this index can be calculated respectively. Among them, the weight percentage of stock A in this index is 31.25%, the weight percentage of stock B in this index is 37.5%, the weight percentage of stock C in this index is 18.75%, and the weight percentage of stock D in this index is 12.5%.

[0055] In some embodiments, for "determine the first similarity matrix based on the constituent stock weight percentages of each ETF and the constituent stock weight percentages of each index", it may specifically include:

[0056] For each ETF, based on the constituent stock weight percentage of this ETF and the constituent stock weight percentages of each index, determine the first similarity result between this ETF and each index;

[0057] Based on the first similarity results between this ETF and each index, determine multiple first similarity results between this ETF and multiple indexes;

[0058] Based on the multiple first similarity results, determine the first similarity matrix.

[0059] Here, for each ETF, based on the weight proportion of one or more stocks in the ETF and the weight proportion of one or more stocks in each index in the corresponding index, the similarity in weight between each stock in the ETF and each stock in each index can be calculated to obtain the first similarity result between the ETF and each index, thereby obtaining multiple first similarity results between the ETF and multiple indexes. And according to the above steps, multiple first similarity results between multiple ETFs and multiple indexes can be further obtained. Based on the multiple first similarity results, a first similarity matrix, that is, an index-ETF similarity matrix, can be constructed. Among them, the first similarity matrix is a symmetric matrix, and the similarity calculation can adopt other calculation methods such as cosine angle, Euclidean distance, Manhattan distance, etc., and it is not limited here.

[0060] For example, if there are m indexes and n ETFs, the component stock weight proportion in each index is expressed as A i =[a 1 ,a 2 ,…,a n , and the component stock weight proportion in each ETF is expressed as B j =[b 1 ,b 2 ,…,b n . Using the cosine similarity calculation method to calculate the similarity between the component stock weight proportion A i of the i-th ETF and the component stock weight proportion B j of the j-th index, the first similarity result s ij between the i-th ETF and the j-th index is obtained, thereby obtaining multiple first similarity results between n ETFs and m indexes, and constructing a first similarity matrix C, whose definition formula is:

[0061]

[0062] Among them, s 11 represents the first similarity result between the 1st ETF and the 1st index, s 1m represents the first similarity result between the 1st ETF and the m-th index, s n1 represents the first similarity result between the n-th ETF and the 1st index, s nm represents the first similarity result between the n-th ETF and the m-th index, i ≤ n, j ≤ m.

[0063] Step 102: Based on the target stock information and the first similarity matrix, determine multiple second similarity results between the target stock and multiple ETFs.

[0064] In an embodiment of the present application, after obtaining the first similarity matrix, to further obtain the similarity between the target stock and the ETF, based on the target stock information and the first similarity matrix, similarity calculations can be performed on the weight proportion of the target stock in each index and the similarity between the constituent stocks of each ETF and each index, to obtain the second similarity results between the target stock and each ETF, thereby determining multiple second similarity results between the target stock and multiple ETFs, so as to comprehensively consider the degree of closeness of the association between the stock and different indexes and ETFs.

[0065] In some embodiments, the target stock information includes the stock code of the target stock; wherein, for step 102, it may specifically include:

[0066] Based on the stock code of the target stock, determine the weight proportion of the target stock in multiple indexes;

[0067] Based on the weight proportion of the target stock in multiple indexes and the first similarity matrix, determine multiple second similarity results.

[0068] Here, first, based on the stock code of the target stock, determine whether each of the multiple indexes includes the target stock. If there is a first index that includes the target stock, then based on the market value of the target stock and the total market value of the first index, determine the weight proportion of the target stock in the first index. If there is a second index that does not include the target stock, then directly determine that the weight proportion of the target stock in the second index is 0, so as to determine the weight proportion of the target stock in multiple indexes. Then, based on the weight proportion of the target stock in multiple indexes and the multiple first similarity results between each ETF included in the first similarity matrix and multiple indexes, calculate the second similarity results between the target stock and each ETF, thereby obtaining multiple second similarity results between the target stock and multiple ETFs.

[0069] In some embodiments, the first similarity matrix includes multiple first similarity results between each ETF among multiple ETFs and multiple indexes; wherein, for "Based on the weight proportion of the target stock in multiple indexes and the first similarity matrix, determine multiple second similarity results", it may specifically include:

[0070] For each ETF, based on the weight proportion of the target stock in multiple indexes and the multiple first similarity results between this ETF and multiple indexes, determine multiple second sub-similarity results between the target stock and this ETF;

[0071] Sum up the multiple second sub-similarity results to obtain the second similarity result between the target stock and this ETF.

[0072] Here, for each ETF, first, based on the weight ratios of the target stocks in multiple indices and multiple first similarity results between the ETF and multiple indices, similarity calculations are performed on the weight ratios of the target stocks in each index and the first similarity results between the ETF and each index to obtain multiple second sub-similarity results between the target stocks and the ETF. Then, the multiple second sub-similarity results are summed to obtain the second similarity result between the target stocks and the ETF. And in accordance with the above steps, multiple second similarity results between the target stocks and multiple ETFs are obtained in sequence.

[0073] For example, if there are m indices and n ETFs, the weight ratios of the target stocks in the m indices are [w 1 , w 2 , …, w m , and the multiple first similarity results between the n ETFs and the m indices are as shown in the above formula (1), then for each ETF, the multiple first similarity results between the i-th ETF and the m indices are [s i1 , s i2 , …, s im . For the weight ratio w j of the target stocks in the j-th index and the first similarity result s ij between the i-th ETF and the j-th index, similarity calculations are performed to obtain the j-th second sub-similarity result w j ×s ij between the target stocks and the i-th ETF. Thus, m second sub-similarity results [w 1 ×s i1 , w 2 ×s i2 , …, w m ×s im between the target stocks and the i-th ETF are obtained. Then, the m second sub-similarity results are summed to obtain the second similarity result S i between the target stocks and the i-th ETF. And in accordance with the above steps, multiple second similarity results [S 1 , S 2 , …, S n between the target stocks and the n ETFs are obtained in sequence.

[0074] Step 103: Based on multiple second similarity results, determine one or more target ETFs and recommend one or more target ETFs.

[0075] In an embodiment of the present application, after obtaining multiple second similarity results between the target stock and multiple ETFs, based on the ETF liquidity information, the comprehensive score corresponding to each ETF can be calculated based on the second similarity result between the target stock and each ETF, and one or more target ETFs can be determined based on the comprehensive score corresponding to each ETF. Then, one or more target ETFs can be recommended as recommended products. Among them, the comprehensive scores corresponding to one or more target ETFs are the highest scores among the multiple comprehensive scores corresponding to the multiple ETFs, which can represent investment characteristics highly similar to the target stock. In this way, it is possible to accurately recommend ETFs that highly match the investment characteristics of the target stock for the target stock, meet the needs of investors for ETFs with highly similar investment characteristics to the stocks they are concerned about, and improve the accuracy of the recommendation results.

[0076] In some embodiments, for "determining one or more target ETFs based on multiple second similarity results" in step 103, it may specifically include:

[0077] Based on multiple second similarity results and the average trading volume of each ETF among the multiple ETFs within a preset time period, determine the multiple comprehensive scores corresponding to the multiple ETFs;

[0078] Sort the multiple comprehensive scores in descending order to obtain the sorted multiple comprehensive scores;

[0079] Select the preset number of ETFs with the highest scores from the sorted multiple comprehensive scores as one or more target ETFs.

[0080] Here, for each ETF, first, based on the second similarity result between the target stock and the ETF and the average trading volume of the ETF within a preset time period, determine the comprehensive score corresponding to the ETF, so as to obtain the multiple comprehensive scores corresponding to the multiple ETFs. Secondly, sort the multiple comprehensive scores in descending order to obtain the sorted multiple comprehensive scores. Then, select the preset number of ETFs with the highest scores from the sorted multiple comprehensive scores as one or more target ETFs, and recommend one or more target ETFs as recommended products.

[0081] It should be noted that since the numerical range of the average trading volume of ETFs may be large, in order to narrow the numerical gap, the average trading volume of ETFs within a preset time period can be numerically transformed (such as logarithmic transformation) first, and then the comprehensive score corresponding to the ETF can be calculated.

[0082] In some embodiments, after determining one or more target ETFs, it may further specifically include:

[0083] Synchronize one or more target ETFs to the database, and create a recommended result data table in the database based on the one or more target ETFs for the user to query the one or more target ETFs based on the recommended result data table.

[0084] Here, the recommended results of one or more target ETFs can be synchronized from the big data warehouse to the database, and a recommended result data table is created in the database based on the one or more target ETFs. Among them, the fields in the table structure of the recommended result database include: target stock code, ETF codes of one or more target ETFs, comprehensive scores corresponding to the one or more target ETFs, and recommended time. Here, the recommended time refers to the timestamp information generated when recording the recommended results.

[0085] It should be noted that after creating the database, the storage capacity and index strategy of the database can also be reasonably planned according to the business volume and data growth expectation to ensure the efficient storage and fast retrieval of data. For example, create indexes for fields such as target stock codes and ETF codes of one or more target ETFs to improve query efficiency to cope with the scenario where users frequently query recommended results.

[0086] Here, when recommending one or more target ETFs, the ETF codes, names, tracked index names, comprehensive scores, and relevant introduction information included in each target ETF can be displayed in a clear, intuitive, and easy-to-operate form in the recommendation section of the stock details page, and a brief recommendation description is provided through floating windows, new user guides, etc., including how to view and utilize these recommended information, and the basis for the recommendation is explained to the user, such as based on the similarity between the target stock and the ETF, the liquidity of the ETF, etc. In this way, it can help users understand the value and usage method of the recommendation service and enhance the user's trust in the recommended results. At the same time, interactive buttons such as "interested", "not interested", "I have questions" can also be set in the recommendation section to facilitate users to instantly feedback their attitudes and questions about the recommended results.

[0087] In the technical solution of the embodiment of the present application, based on the constituent stock information of multiple exchange-traded funds (ETFs) and the constituent stock information of multiple indexes, a first similarity matrix between the multiple ETFs and the multiple indexes is determined. Based on the target stock information and the first similarity matrix, multiple second similarity results between the target stock and the multiple ETFs are determined. Based on the multiple second similarity results, one or more target ETFs are determined, and one or more target ETFs are recommended. In this way, by analyzing the similarity of the constituent stocks between ETF funds and each index and integrating it into the evaluation system of the overall similarity between the target stock and the ETF, it can be ensured that the recommended ETFs have a strong correlation with the target stock in terms of constituent stocks, so as to accurately recommend ETF funds that highly match the investment characteristics of the target stock, meet the needs of investors for ETFs with highly similar investment characteristics to the stocks they are concerned about, and improve the accuracy and effectiveness of ETF recommendation.

[0088] The embodiment of the present application also proposes an intelligent ETF recommendation method for stocks. Figure 2 It is a schematic flowchart of the intelligent ETF recommendation method for stocks provided by the embodiment of the present application, as Figure 2 shown. The method includes the following steps:

[0089] Step 201: Obtain relevant information data based on an external information data source.

[0090] Here, by establishing a data connection channel with the external information data source and through corresponding data interfaces and synchronization tools, information data such as Wind is synchronized to the big data warehouse according to specific rules and frequencies, so as to extract various basic data in the unified data storage and management environment of the big data warehouse.

[0091] Step 202: Perform data extraction on the relevant information data to obtain extracted data.

[0092] Specifically, obtain the basic information of stocks in the big data warehouse, including the stock code and name, which can uniquely identify each stock and facilitate subsequent analysis and processing. Similarly, obtain the basic information of ETFs, including the ETF code and name, to distinguish different ETF products. Then, according to specific business requirements, select a part of the ETFs to form a product pool. For example, the ETFs that the company focuses on selling can be included in the product pool because these ETFs are often verified by the market, have certain advantages, and meet the key points of the company's business promotion. Or select the ETFs with relatively high trading volumes in the recent period. A high trading volume usually means a high level of activity and attention in the market, relatively good liquidity, and is more likely to become an investment target favored by investors.

[0093] In addition, it is also possible to obtain the constituent stock information of various indices, including the index code, constituent stock code, and the proportion of constituent stocks. These indices cover a variety of representative indices such as Shenwan industry indices, A-share indices, and Hong Kong stock indices. The index code is used to identify different indices, the constituent stock code is used to determine which stocks constitute the corresponding index, and further calculate the proportion of each constituent stock in the corresponding index. Among them, when calculating the proportion of constituent stocks, the latest data needs to be used to ensure that the weight proportion of each stock reflected in the index conforms to the actual situation of the current market.

[0094] After obtaining the basic information of the ETF, it is also necessary to obtain the constituent stock information of the ETF, including the ETF code, constituent stock code, and the proportion of constituent stocks. Among them, considering the timeliness and effectiveness of the data, the publicly available data for the latest quarter can be obtained. Since the portfolio adjustment of the ETF usually has a certain cycle, the data for one quarter can not only reflect its relatively stable constituent stock composition but also better reflect the recent change trend.

[0095] Step 203: Based on the extracted data, calculate the similarity of the constituent stocks of the index and the ETF to obtain an index-ETF similarity matrix. Based on the index-ETF similarity matrix and the weight proportion of the stocks in each index, calculate the weighted similarity of the stocks and the ETF to obtain multiple similarity results of the stocks and the ETF.

[0096] First, based on the constituent stock information of the index and the constituent stock information of the ETF, calculate the similarity of the constituent stocks of the index and the ETF. Assume that vector A and vector B represent the constituent stock information of the index and the ETF respectively, and the elements in the vector are the weight proportions of the corresponding constituent stocks. For two n-dimensional vectors A = [a 1 , a 2 , …, a n and B = [b 1 , b 2 , …, b n , the calculation formula for the cosine similarity between the two is:

[0097]

[0098] Among them, θ represents the angle between vector A and vector B. Since the elements in the vector are the weight proportions of the corresponding constituent stocks, they are all non-negative. Therefore, the value range of cos(θ) is between [0, 1]. Among them, the closer the value of cos(θ) is to 1, the higher the similarity between vector A and vector B, that is, the more consistent the constituent stock situations of the index and the ETF; the closer the value of cos(θ) is to 0, the lower the similarity between vector A and vector B, that is, the index and the ETF basically have no overlapping constituent stocks.

[0099] Calculate the similarity of the constituent stocks of each index and each ETF through the above formula (2), and multiple constituent stock similarity results can be obtained. Through these multiple constituent stock similarity results, an index-ETF similarity matrix can be finally obtained, which clearly shows the similarity degree relationship of each index and different ETFs in terms of constituent stocks.

[0100] After calculating the similarity of the constituent stocks of each index and each ETF, based on the calculated index-ETF similarity matrix and the weight proportion of the stock in each index, the similarity of the stock and each ETF can be further calculated.

[0101] Suppose there are m indexes and n ETFs, and the weight proportion of the stock in the i-th index is denoted as w i (i = 1, 2,..., m), and the similarity of the constituent stocks of the i-th index and the j-th ETF (obtained through the previously calculated cosine similarity) is denoted as s ij , then the similarity S j of the stock and the j-th ETF

[0102]

[0103] The above formula (3) means that multiply the weight proportion of the stock in the i-th index by the similarity of the constituent stocks of the j-th ETF and the i-th index, and then sum to obtain the similarity of the stock and the j-th ETF, thereby comprehensively considering the degree of closeness of the stock in different indexes and ETFs, and finally obtaining the similarity results of the stock and each ETF.

[0104] Step 204: Calculate the comprehensive score of each ETF based on the similarity results of the stock and each ETF, and sort the comprehensive scores of each ETF in descending order, and select several ETFs with the highest scores.

[0105] Here, on the basis of considering the ETF liquidity information, calculate the comprehensive score of each ETF based on the similarity of the stock and each ETF, and sort in descending order to determine the final list of ETFs recommended to investors. Suppose T j is the average turnover of the j-th ETF in the past 5 trading days, and S j is the similarity result of the stock and the j-th ETF calculated above. Then the comprehensive score C j of the j-th ETF

[0106] C j = ln(T j ) × S j (4)

[0107] Among them, for ln(T j), the role of the natural logarithm function is to calculate the average trading volume T of the j-th ETF in the past 5 trading days j for numerical transformation. Since the numerical range of the trading volume may be large, taking the logarithm can, to a certain extent, narrow the numerical gap, making the influence of the liquidity factors of different ETFs in the comprehensive score more reasonable.

[0108] After calculating the comprehensive score of each ETF, a score list [C 1 , C 2 , …, C j can be obtained. Sort this score list in descending order, and select several of the highest-ranked ETFs (usually 3) from the sorted score list as candidate recommended ETF products.

[0109] Step 205: Store the ETF recommendation results in a relational database, and reasonably plan the storage capacity and indexing strategy of the database.

[0110] Synchronize the recommendation results from the big data warehouse to the relational database, and create a recommendation result data table. Among them, the fields included in the table structure of the recommendation result data table are stock code (the unique identifier of the target stock corresponding to the recommended ETF), ETF code (the recommended ETF), comprehensive score (used to measure the priority ranking of this ETF in this recommendation), and recommendation time (recording the timestamp when the recommendation result is generated, which is convenient for subsequent tracking and analysis of the recommendation situation in different periods). At the same time, according to the business volume and data growth expectations, reasonably plan the storage capacity and indexing strategy of the database to ensure the efficient storage and quick retrieval of data. For example, establish indexes for fields such as stock code and ETF code that are commonly used as query conditions to improve the query efficiency to cope with the scenario where investors frequently query the recommendation results.

[0111] Step 206: Display the ETF recommendation results in the ETF recommendation module on the stock details page, and improve the user experience through functions such as novice instructions, interaction, and feedback when displaying the ETF recommendation results for users.

[0112] Display the list of recommended ETFs in a clear, intuitive, and easy-to-operate form in a prominent position on the stock details page, such as the sidebar on the right side of the page, the "Related Recommendations" section below, etc. Each displayed item of the recommended ETF includes the ETF code, name, name of the tracked index, comprehensive score, and a short highlight introduction (such as features and advantages like low fee rate, high dividend rate, etc.). At the same time, set obvious clickable links to facilitate users to directly jump to the detailed introduction page of the ETF to further understand the details.

[0113] In addition, when presenting ETF recommendation results to users for the first time, provide a brief recommendation explanation through methods such as floating windows and new user guides, explaining to users the basis for the recommendation (such as factors based on the similarity between stocks and ETFs, the liquidity of ETFs, etc.) and how to view and utilize this recommended information, helping users understand the value and usage method of the recommendation service, and enhancing users' trust in the recommendation results. At the same time, set interactive buttons in the display area of the recommended ETFs, such as "interested", "not interested", "I have questions", etc., to facilitate users to immediately feedback their attitudes and questions regarding the recommendation results.

[0114] In the technical solution of the embodiment of the present application, by implementing a set of precise and market-actual situation-compliant processes from data acquisition, data extraction, similarity calculation, sorting and screening, storage and query to final operation and promotion, it effectively makes up for the deficiencies in recommending ETFs for stocks in related technologies, and provides a more valuable investment reference service for investors. Among them, based on the defects existing in related technologies, the embodiment of the present application can solve the following technical problems:

[0115] (1) Improve the accuracy of ETF recommendations. Related technologies only rely on industry classification to screen and recommend ETFs, without fully considering key factors such as the component ratio of stocks in each index and the similarity between the constituent stocks of ETFs and the index, resulting in a weak correlation and poor matching degree between the recommended ETFs and stocks. The technical solution of the present application aims to comprehensively consider the above key elements, use a scientific and reasonable weighted calculation method to accurately measure the similarity between stocks and ETFs, so as to accurately recommend ETFs that highly match the investment characteristics of each stock, effectively meet the needs of investors for ETFs with highly similar investment characteristics to the stocks they are concerned about, improve the accuracy of the recommendation results, and help investors make investment decisions that better meet their own expectations.

[0116] (2) Deepen the consideration of the characteristics of constituent stocks. Related technologies lack in-depth exploration of the similarity between the constituent stocks of ETF funds and the corresponding index, resulting in a weak correlation between the actual constituent stocks of the recommended ETFs and the target stocks although they track the same industry index, and unable to provide effective investment references for investors. The technical solution of the present application focuses on solving this problem, deeply analyzing the similarity between the constituent stocks of ETFs and the index during the calculation process, and integrating it into the evaluation system of the overall similarity between stocks and ETFs, ensuring that the recommended ETFs have a strong correlation with the target stocks at the constituent stock level, thereby enhancing the effectiveness and practicality of the recommendation, and providing more valuable investment reference information for investors.

[0117] (3) Optimize the singularity of the sorting basis. When determining the recommendation order in related technologies, most rely on a single factor such as asset size or average daily trading volume for sorting, ignoring other important features of ETFs and unable to comprehensively evaluate the matching degree between ETFs and stocks. The technical solution of this application aims to solve this problem. By constructing a sorting mechanism that comprehensively considers multi-dimensional factors, in addition to considering the similarity between stocks and ETFs, ETF liquidity information is also incorporated (such as multiplying ln(average trading volume of the ETF in the last 5 trading days) and the similarity between the stock and the ETF to obtain a comprehensive score), and all aspects of factors are comprehensively weighed to sort the recommendation results, so that the finally recommended ETFs are selected from the perspective of multi-dimensional matching and are the most suitable choices for investors.

[0118] An intelligent recommendation device applied to a big data warehouse is also proposed in an embodiment of this application. Figure 3 It is a schematic structural diagram of the intelligent recommendation device applied to the big data warehouse provided by the embodiment of this application, as Figure 3 shown. The device includes:

[0119] A first determination unit 301, configured to determine a first similarity matrix between multiple exchange-traded funds (ETFs) and multiple indexes based on the constituent stock information of multiple ETFs and the constituent stock information of multiple indexes.

[0120] A second determination unit 302, configured to determine multiple second similarity results between a target stock and multiple ETFs based on the target stock information and the first similarity matrix.

[0121] A third determination unit 303, configured to determine one or more target ETFs based on multiple second similarity results.

[0122] A recommendation unit 304, configured to recommend one or more target ETFs.

[0123] In some embodiments, the constituent stock information of multiple ETFs includes the ETF code and constituent stock code of each ETF, and the constituent stock information of multiple indexes includes the index code and constituent stock code of each index; wherein,

[0124] The first determination unit 301 is specifically configured to: determine the constituent stock weight ratio of each ETF based on the ETF code and constituent stock code of each ETF; determine the constituent stock weight ratio of each index based on the index code and constituent stock code of each index; and determine the first similarity matrix based on the constituent stock weight ratio of each ETF and the constituent stock weight ratio of each index.

[0125] In some embodiments, the first determination unit 301 is further specifically configured to: determine one or more stocks included in each index based on the index code and constituent stock code of each index; calculate the weight ratio of each stock in each index for each of the one or more stocks to obtain the constituent stock weight ratio of each index.

[0126] In some embodiments, the first determination unit 301 is further specifically configured to: for each ETF, determine a first similarity result between the ETF and each index based on the constituent stock weight ratio of the ETF and the constituent stock weight ratio of each index; determine multiple first similarity results between the ETF and multiple indexes based on the first similarity results between the ETF and each index; determine a first similarity matrix based on the multiple first similarity results.

[0127] In some embodiments, the target stock information includes the stock code of the target stock; wherein,

[0128] The second determination unit 302 is specifically configured to: determine the weight ratio of the target stock in multiple indexes based on the stock code of the target stock; determine multiple second similarity results based on the weight ratio of the target stock in multiple indexes and the first similarity matrix.

[0129] In some embodiments, the first similarity matrix includes multiple first similarity results between each of multiple ETFs and multiple indexes; wherein,

[0130] The second determination unit 302 is further specifically configured to: for each ETF, determine multiple second sub-similarity results between the target stock and the ETF based on the weight ratio of the target stock in multiple indexes and the multiple first similarity results between the ETF and multiple indexes; sum the multiple second sub-similarity results to obtain the second similarity result between the target stock and the ETF.

[0131] In some embodiments, the third determination unit 303 is specifically configured to: determine multiple comprehensive scores corresponding to multiple ETFs based on the multiple second similarity results and the average trading volume of each ETF in multiple ETFs within a preset time period; sort the multiple comprehensive scores in descending order to obtain the sorted multiple comprehensive scores; select a preset number of ETFs with the highest scores from the sorted multiple comprehensive scores as one or more target ETFs.

[0132] In some embodiments, the apparatus further includes: a synchronization unit; wherein,

[0133] A synchronization unit is used to synchronize one or more target ETFs to a database, and create a recommended result data table based on the one or more target ETFs in the database for users to query the one or more target ETFs based on the recommended result data table.

[0134] In the technical solution of the embodiment of the present application, based on the constituent stock information of multiple Exchange Traded Funds (ETFs) and the constituent stock information of multiple indexes, a first similarity matrix between the multiple ETFs and the multiple indexes is determined. Based on the target stock information and the first similarity matrix, multiple second similarity results between the target stock and the multiple ETFs are determined. Based on the multiple second similarity results, one or more target ETFs are determined and recommended for the one or more target ETFs. In this way, by analyzing the similarity of the constituent stocks between the ETF funds and each index and integrating it into the evaluation system of the overall similarity between the target stock and the ETF, it can be ensured that the recommended ETFs have a strong correlation with the target stock in terms of constituent stocks, so as to accurately recommend ETF funds that highly match the investment characteristics of the target stock, meet the needs of investors for ETFs with highly similar investment characteristics to the stocks they are concerned about, and improve the accuracy and effectiveness of ETF recommendations.

[0135] Those skilled in the art should understand that Figure 3 The implementation functions of the units in the intelligent recommendation device shown can be understood with reference to the relevant descriptions of the foregoing method. Figure 3 The functions of the units in the intelligent recommendation device shown can be implemented by a program running on a processor or by specific logic circuits.

[0136] Figure 4 is a schematic structural diagram of a processing device provided by an embodiment of the present application. The processing device can be a terminal device or a network device. Figure 4 The processing device shown includes a processor 401, and the processor 401 can call and run a computer program from a memory to implement the method in the embodiment of the present application.

[0137] Optionally, as Figure 4 shown, the processing device may further include a memory 402. Among them, the processor 401 can call and run a computer program from the memory 402 to implement the method in the embodiment of the present application.

[0138] Among them, the memory 402 can be a separate device independent of the processor 401 or integrated in the processor 401.

[0139] Optionally, as Figure 4As shown, the processing device may further include a transceiver 403. The processor 401 may control the transceiver 403 to communicate with other devices. Specifically, it may send information or data to other devices, or receive information or data sent by other devices.

[0140] Among them, the transceiver 403 may include a transmitter and a receiver. The transceiver 403 may further include an antenna, and the number of antennas may be one or more.

[0141] Specifically, the processing device may be the intelligent recommendation device of the embodiment of the present application, and the processing device may implement the corresponding processes implemented by the various methods of the embodiment of the present application. For the sake of brevity, details are not described herein again.

[0142] It should be understood that the processor of the embodiment of the present application may be an integrated circuit chip with signal processing capabilities. In the implementation process, the steps of the above method embodiments may be completed by the integrated logic circuit in the hardware of the processor or by instructions in the form of software. The above processor may be a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It may implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application may be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method.

[0143] It can be understood that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM), and direct rambus random access memory (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include but not be limited to these and any other suitable types of memory.

[0144] It should be understood that the above memory is by way of example but not limitation. For example, the memory in the embodiments of the present application can also be a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate SDRAM (DDR SDRAM), an enhanced SDRAM (ESDRAM), a synch link DRAM (SLDRAM), and a direct rambus random access memory (DR RAM), etc. That is to say, the memory in the embodiments of the present application is intended to include but not be limited to these and any other suitable types of memory.

[0145] The embodiments of the present application also provide a computer-readable storage medium for storing a computer program. The computer-readable storage medium can be applied to the processing device in the embodiments of the present application, and the computer program enables the computer to execute the corresponding processes implemented by the various methods in the embodiments of the present application. For the sake of brevity, it will not be described in detail here.

[0146] The embodiments of the present application also provide a computer program product including computer program instructions. The computer program product can be applied to the processing device in the embodiments of the present application, and the computer program instructions enable the computer to execute the corresponding processes implemented by the various methods in the embodiments of the present application. For the sake of brevity, it will not be described in detail here.

[0147] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0148] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be described in detail here.

[0149] In several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling, direct coupling, or communication connection can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be in an electrical, mechanical, or other form.

[0150] 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 can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0151] In addition, the functional units in the various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0152] When the above-mentioned functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0153] As described above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. An intelligent recommendation method, characterized in that: Applied to a big data warehouse, the method includes: Based on the constituent stock information of a plurality of exchange-traded open-end index funds (ETFs) and the constituent stock information of a plurality of indices, determining a first similarity matrix between the plurality of ETFs and the plurality of indices; Based on the target stock information and the first similarity matrix, determining a plurality of second similarity results between the target stock and the plurality of ETFs; Based on the multiple second similarity results, one or more target ETFs are determined, and the one or more target ETFs are recommended.

2. The method according to claim 1, characterized in that The constituent stock information of the plurality of ETFs includes the ETF code and constituent stock code of each ETF, and the constituent stock information of the plurality of indices includes the index code and constituent stock code of each index; The determining, based on the constituent stock information of the plurality of ETFs and the constituent stock information of the plurality of indices, a first similarity matrix between the plurality of ETFs and the plurality of indices comprises: Determine the weight ratio of the constituent stocks of each ETF based on the ETF code and constituent stock code of each ETF; Determine the weight ratio of the constituent stocks of each index based on the index code and constituent stock code of each index; The first similarity matrix is ​​determined based on the weight ratios of the constituent stocks of each ETF and the weight ratios of the constituent stocks of each index.

3. The method according to claim 2, characterized in that The determining the weight ratio of the constituent stocks of each index based on the index code and the constituent stock code of each index includes: Determine one or more stocks included in each index based on the index code and constituent stock codes of each index; Based on each stock in the one or more stocks, the weight ratio of each stock in each index is calculated to obtain the weight ratio of the constituent stocks of each index.

4. The method according to claim 2, characterized in that: The determining the first similarity matrix based on the weight ratio of the constituent stocks of each ETF and the weight ratio of the constituent stocks of each index includes: For each of the ETFs, based on the weight ratios of the constituent stocks of the ETF and the weight ratios of the constituent stocks of each of the indexes, determine a first similarity result between the ETF and each of the indexes; Determining a plurality of first similarity results between the ETF and the plurality of indices based on the first similarity result between the ETF and each of the indices; Based on the plurality of first similarity results, the first similarity matrix is ​​determined.

5. The method according to claim 1, characterized in that The target stock information includes the stock code of the target stock; The step of determining a plurality of second similarity results between the target stock and the plurality of ETFs based on the target stock information and the first similarity matrix includes: Based on the stock code of the target stock, determining the weight ratio of the target stock in the multiple indexes; The plurality of second similarity results are determined based on the weight ratio of the target stock in the plurality of indexes and the first similarity matrix.

6. The method according to claim 5, characterized in that The first similarity matrix includes a plurality of first similarity results between each ETF in the plurality of ETFs and the plurality of indices; The determining the plurality of second similarity results based on the weight ratio of the target stock in the plurality of indexes and the first similarity matrix includes: For each of the ETFs, based on the weight ratio of the target stock in the multiple indexes and the multiple first similarity results between the ETF and the multiple indexes, determine multiple second sub-similarity results between the target stock and the ETF; The multiple second sub-similarity results are summed to obtain a second similarity result between the target stock and the ETF.

7. The method according to claim 1, characterized in that The step of determining one or more target ETFs based on the plurality of second similarity results comprises: Determine a plurality of comprehensive scores corresponding to the plurality of ETFs based on the plurality of second similarity results and an average transaction amount of each of the plurality of ETFs within a preset time period; Sorting the multiple comprehensive scores in descending order to obtain a plurality of sorted comprehensive scores; A preset number of ETFs with the highest scores are selected from the sorted multiple comprehensive scores as the one or more target ETFs.

8. The method according to any one of claims 1 to 7, characterized in that The method further comprises: The one or more target ETFs are synchronized to a database, and a recommendation result data table is created in the database based on the one or more target ETFs, so that a user can query the one or more target ETFs based on the recommendation result data table.

9. An intelligent recommendation device, characterized in that: Applied to a big data warehouse, the device comprises: A first determining unit is used to determine a first similarity matrix between the plurality of exchange-traded open-end index funds (ETFs) and the plurality of indexes based on the constituent stock information of the plurality of ETFs and the constituent stock information of the plurality of indexes; A second determining unit, configured to determine a plurality of second similarity results between the target stock and the plurality of ETFs based on the target stock information and the first similarity matrix; A third determining unit, configured to determine one or more target ETFs based on the plurality of second similarity results; A recommendation unit is used to recommend the one or more target ETFs.

10. A processing device, characterized in that: include: A processor and a memory, the memory being used to store a computer program, the processor being used to call and run the computer program stored in the memory to execute the method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: Used to store a computer program, wherein the computer program causes a computer to execute the method according to any one of claims 1 to 8.

12. A computer program product, characterized in that The method comprises computer program instructions which cause a computer to execute the method as claimed in any one of claims 1 to 8.

Images