Market emotion index construction method and device, computer equipment and storage medium

Abstract

The invention relates to a market emotion index construction method and device, computer equipment and a storage medium. The method comprises the steps that risk index data corresponding to a preset historical time interval are acquired, the risk index data comprise actual resource numerical values corresponding to unit time in the historical time interval, then multiple pieces of effective market emotion factor data corresponding to the historical time interval are acquired according to the risk index data, and the effective market emotion factor data are stored in the historical time interval; the correlation between the effective market emotion factor data and the actual resource value meets a preset condition, and then a market emotion index is constructed according to the risk index data and the effective market emotion factor data. By adopting the method, the accuracy of market emotion index construction can be improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, apparatus, computer equipment, and storage medium for constructing a market sentiment index. Background Art

[0002] With the development of the economy, financial markets have attracted widespread attention, and monitoring the risk index of financial markets is of great significance to promoting financial stability.

[0003] Risk indices are usually affected by multidimensional factors, such as the market sentiment index of the financial market. The market sentiment index is constructed by integrating market sentiment data from various financial markets, such as the daily return rate of stocks and the number of stock gains and losses. Traditional market sentiment index construction methods usually involve manually selecting multiple market sentiment data for unsupervised learning.

[0004] However, the above-mentioned method for constructing the market sentiment index has the problem of low construction accuracy. Summary of the Invention

[0005] Based on this, it is necessary to provide a market sentiment index construction method, device, computer equipment and storage medium that can improve the construction accuracy in response to the above technical problems.

[0006] In a first aspect, the present application provides a method for constructing a market sentiment index. The method comprises:

[0007] Obtain risk indicator data corresponding to a preset historical time interval, including actual resource values ​​corresponding to each unit time within the historical time interval;

[0008] Based on the risk indicator data, multiple valid market sentiment factor data corresponding to the historical time interval are obtained, and the correlation between the valid market sentiment factor data and the actual resource value meets the preset conditions;

[0009] A market sentiment index is constructed based on risk indicator data and data of various effective market sentiment factors.

[0010] In one embodiment, a market sentiment index is constructed based on risk indicator data and data of various effective market sentiment factors, including:

[0011] Obtain the target correlation coefficient corresponding to each effective market sentiment factor data based on each effective market sentiment factor data and risk indicator data;

[0012] The market sentiment index is constructed based on the target correlation coefficients and the effective market sentiment factor data.

[0013] In one embodiment, a market sentiment index is constructed based on each target correlation coefficient and each effective market sentiment factor data, including:

[0014] The target correlation coefficients are used as factor weights, and the effective market sentiment factor data are statistically processed according to the factor weights to obtain the market sentiment index.

[0015] In one embodiment, obtaining a target correlation coefficient corresponding to each effective market sentiment factor data based on each effective market sentiment factor data and risk indicator data includes:

[0016] According to the time series, the data of each effective market sentiment factor is divided into training factor data and testing factor data;

[0017] According to each training factor data and risk indicator data, the initial correlation coefficient corresponding to each training factor data is obtained;

[0018] The test sentiment index is constructed using the initial correlation coefficients and the test factor data, and the validity of the test sentiment index is verified based on the risk indicator data;

[0019] If the verification is successful, the initial correlation coefficients are used as the target correlation coefficients.

[0020] In one embodiment, based on the risk indicator data, multiple valid market sentiment factor data corresponding to the historical time interval are obtained, including:

[0021] Acquire multiple initial market sentiment factor data corresponding to a historical time interval, and extract sentiment trend item data and sentiment detrended item data corresponding to the initial market sentiment factor data based on the initial market sentiment factor data;

[0022] Preprocessing each emotion trend item data and each emotion detrended item data to obtain a plurality of candidate trend item data and a plurality of candidate detrended item data;

[0023] Based on the risk indicator data, the candidate trend item data and the candidate detrended item data are screened and processed to obtain multiple effective market sentiment factor data.

[0024] In one embodiment, each candidate trend item data and each candidate detrended item data are screened based on the risk indicator data to obtain multiple valid market sentiment factor data, including:

[0025] Based on the risk indicator data, single factor screening is performed on each candidate trend item data and each candidate detrended item data to obtain multiple single factor data. The single factor screening process includes stationarity test processing, correlation analysis processing, linear regression test processing and causal test processing;

[0026] Based on the risk indicator data, each single factor data is subjected to multi-factor screening processing to obtain multiple effective market sentiment factor data. The multi-factor screening processing includes combined linear test processing.

[0027] In a second aspect, the present application also provides a device for constructing a market sentiment index. The device comprises:

[0028] A first acquisition module is used to obtain risk indicator data corresponding to a preset historical time interval, where the risk indicator data includes actual resource values ​​corresponding to each unit time in the historical time interval;

[0029] The second acquisition module is used to obtain multiple valid market sentiment factor data corresponding to the historical time interval based on the risk indicator data, and the correlation between the valid market sentiment factor data and the actual resource value meets the preset conditions;

[0030] The construction module is used to construct a market sentiment index based on risk indicator data and data of various effective market sentiment factors.

[0031] In a third aspect, an embodiment of the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method of the first aspect described above when executing the computer program.

[0032] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method of the first aspect described above when the computer program is executed by a processor.

[0033] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the steps of the method in the first aspect.

[0034] The above-mentioned market sentiment index construction method, device, computer equipment and storage medium obtain risk indicator data corresponding to a preset historical time interval, the risk indicator data includes the actual resource value corresponding to each unit time in the historical time interval, and then, based on the risk indicator data, obtains multiple valid market sentiment factor data corresponding to the historical time interval, the correlation between the valid market sentiment factor data and the actual resource value meets the preset conditions, and then constructs the market sentiment index based on the risk indicator data and each valid market sentiment factor data. In this way, the multiple sentiment factor data in the historical time interval are screened by the acquired risk indicator data, and multiple valid market sentiment factor data whose correlation with the risk indicator data meets the preset conditions are obtained, thereby avoiding the problem of inaccurate selection of sentiment data in traditional technologies through manual selection. In addition, the embodiment of the present application constructs a market sentiment index based on valid market sentiment factor data and risk indicator data, thereby avoiding the problem of inaccurate market sentiment index construction in the unsupervised sentiment index construction method in traditional technologies, and improving the accuracy of market sentiment index construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 This is a diagram of an application environment of a method for constructing a market sentiment index in one embodiment;

[0037] Figure 2 A flowchart of a method for constructing a market sentiment index in one embodiment is shown;

[0038] Figure 3 is a schematic flow chart of step 203 in another embodiment;

[0039] Figure 4 301 is a flow chart of another embodiment;

[0040] Figure 5 A schematic diagram of validating the testing of sentiment index and risk indicator data in another embodiment;

[0041] Figure 6 202 is a flow chart of step 202 in another embodiment;

[0042] Figure 7 is a schematic flow chart of step 503 in another embodiment;

[0043] Figure 8 A flowchart of a method for constructing a market sentiment index in one embodiment is shown;

[0044] Figure 9 A schematic diagram of a single factor screening process in one embodiment;

[0045] Figure 10 A schematic diagram of a multi-factor screening process in one embodiment;

[0046] Figure 11 This is a structural block diagram of a market sentiment index construction device in one embodiment;

[0047] Figure 12 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0049] With the continuous development of international financial markets and the increasing number of financial risks, financial risk monitoring and early warning have become a global research hotspot. Establishing effective financial risk monitoring and early warning methods will help improve the scientific nature and effectiveness of financial risk management, enhance the transparency and stability of financial markets, and thus have important implications for promoting financial stability.

[0050] Traditional risk monitoring and early warning methods are usually based on the analysis of fundamental factors such as financial data and economic data to obtain risk indicator data for a specific time period. However, financial data and economic data are published annually or monthly, which has a lag and low data frequency, so the accuracy and practicality of this risk monitoring and early warning method need to be improved.

[0051] Therefore, the market sentiment index has been introduced as one of the indicators for monitoring financial risk. This index is constructed by integrating market sentiment data from various financial markets. Market sentiment data reflects investors' expectations and confidence in the market. Investors' emotions and psychological state can influence the market, thereby influencing stock prices and volatility. Therefore, the market sentiment index constructed from multidimensional market sentiment data is also a key factor influencing risk indicators. Market sentiment data comes in a variety of types, such as daily stock returns and the number of stocks that rise and fall.

[0052] Traditional methods of constructing market sentiment indexes usually involve manually selecting a variety of market sentiment data for unsupervised learning, which results in low construction accuracy.

[0053] In view of this, the embodiments of the present application provide a method, apparatus, computer equipment and storage medium for constructing a market sentiment index, which obtains risk indicator data corresponding to a preset historical time interval, the risk indicator data including the actual resource values ​​corresponding to each unit time in the historical time interval, and then obtains multiple valid market sentiment factor data corresponding to the historical time interval based on the risk indicator data, the correlation between the valid market sentiment factor data and the actual resource values ​​meeting the preset conditions, and then constructs a market sentiment index based on the risk indicator data and the valid market sentiment factor data. In this way, by using the acquired risk indicator data, the multiple sentiment factor data in the historical time interval are screened, and multiple valid market sentiment factor data whose correlation with the risk indicator data meets the preset conditions are obtained, thereby avoiding the problem of inaccurate selection of sentiment data in the traditional technology by manually selecting sentiment data, and constructing the market sentiment index based on the valid market sentiment factor data and the risk indicator data, thereby avoiding the problem of inaccurate market sentiment index construction in the unsupervised sentiment index construction method in the traditional technology, and improving the accuracy of market sentiment index construction.

[0054] The market sentiment index construction method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown, the data storage system can store data that the server 102 needs to process. The data storage system can be integrated on the server 102, or placed on the cloud or other network servers. The server 102 can be implemented as an independent server or a server cluster consisting of multiple servers.

[0055] In an exemplary embodiment, Figure 2 As shown in the figure, a method for constructing a market sentiment index is provided, which is applied to Figure 1 Taking the server 102 in FIG. 1 as an example, the method includes the following steps 201 to 203. In which:

[0056] Step 201: Obtain risk indicator data corresponding to a preset historical time interval.

[0057] Risk indicator data is a comprehensive index used to characterize the risk profile of financial markets. It includes actual resource values ​​corresponding to each unit of time within a historical time interval. In one possible implementation, the risk indicator data is daily data, with the unit of time being "day." The actual resource value can be the daily closing price of a specific stock or a comprehensive securities index. For example, the actual resource value is the daily closing price of the CSI 300 Index.

[0058] In one possible implementation, the server obtains the risk indicator data corresponding to the historical time interval input by the user through other devices. In another possible implementation, after the server obtains the actual resource values ​​corresponding to each unit time in the historical time interval, it preprocesses the actual resource values ​​to obtain the risk indicator data corresponding to the historical time interval.

[0059] Regarding the method for the server to preprocess the actual resource values, in one possible implementation, in order to eliminate the nonlinear trend of the actual resource values, the server performs a logarithmic transformation on the actual resource values ​​corresponding to each unit time in the historical time interval. For example, if the actual resource value on day t is a, the value on day t is logarithmically transformed to lna, and lna replaces the original actual resource value a. In another possible implementation, in order to make the risk indicator data more stable in time series, the server performs a detrending operation on the actual resource values ​​corresponding to each unit time in the historical time interval, thereby obtaining detrended item data of the actual resource values ​​corresponding to each unit time, and then uses the detrended item data as risk indicator data.

[0060] Regarding the method for a server to extract detrended data for actual resource values ​​corresponding to each unit time, taking the actual resource value being the daily closing price of the CSI 300 Index as an example, the server obtains the daily closing price of the CSI 300 Index within a preset historical time interval. First, the server extracts trend data for the daily closing price of the CSI 300 Index. Regarding the extraction of trend data, the server presets a moving period based on the application scenario, which ranges from 20 to 100 days. For a closing price of day t, the server obtains the closing prices of multiple days within the moving period before and after time t. For example, if the moving period is 20 days, the server obtains closing price data for a total of 41 days, including the 20 days before and after day t, calculates the average of the 41-day closing price data, and then subtracts the average of the 41-day closing price data from the closing price of day t to obtain the trend data for day t. The server then subtracts the corresponding trend data for each day from the closing price of each day to obtain the detrended data for the daily closing price of the CSI 300 Index.

[0061] Step 202: Acquire multiple valid market sentiment factor data corresponding to historical time intervals based on the risk indicator data.

[0062] The effective market sentiment factor data is the market sentiment factor data that is screened by the server from multiple candidate market sentiment factor data based on the risk indicator data and has a correlation with the risk indicator data that meets the preset conditions. The correlation between the effective market sentiment factor data and the actual resource value meets the preset conditions.

[0063] Market sentiment factor data is generated by preprocessing market sentiment data from multiple categories on the server. Market sentiment data reflects investors' expectations and confidence in the market. Investors' emotions and psychological states can influence the market, thereby impacting stock prices and volatility. Therefore, the market sentiment index constructed from multidimensional market sentiment data is also a key factor influencing risk indicator data. Market sentiment data comes in many forms, such as daily stock returns and the number of stock gains and losses. Optionally, market sentiment factor data represents trend data within a specific category of market sentiment data; alternatively, market sentiment factor data represents non-trend data within a specific category of market sentiment data.

[0064] In a possible implementation, the candidate market sentiment factor data is obtained after the server preprocesses the market sentiment factor data. The preprocessing methods used by the server to preprocess the market sentiment factor data include stationarity processing, numerical normalization processing, etc.

[0065] Regarding the method in which the server screens and processes multiple candidate market sentiment factor data according to preset conditions to obtain effective market sentiment factor data, in one possible implementation manner, the server performs single-factor screening processing on each candidate market sentiment factor data, and then performs multi-factor screening processing on each market sentiment factor data. The single-factor screening processing method includes stationarity test screening, correlation test screening, linear relationship test screening, and Granger causality test screening, etc. The multi-factor screening processing method includes combined linear relationship test screening, etc.

[0066] Step 203: construct a market sentiment index based on the risk indicator data and the data of each effective market sentiment factor.

[0067] After obtaining multiple valid market sentiment factor data, in one possible implementation, the server performs correlation analysis on each valid market sentiment factor data based on the risk indicator data, obtains the correlation coefficient corresponding to each valid market sentiment factor data, and then constructs a market sentiment index based on the correlation coefficient corresponding to each valid market sentiment factor data and each valid market sentiment factor data; in another possible implementation, the server uses each valid market sentiment factor data and risk indicator data as sample data, constructs a market sentiment index network model according to certain constraints, and thus constructs a market sentiment index.

[0068] The above-mentioned market sentiment index construction method obtains risk indicator data corresponding to a preset historical time interval. The risk indicator data includes the actual resource values ​​corresponding to each unit time in the historical time interval. Then, based on the risk indicator data, multiple valid market sentiment factor data corresponding to the historical time interval are obtained. The correlation between the valid market sentiment factor data and the actual resource values ​​meets the preset conditions. Then, based on the risk indicator data and the valid market sentiment factor data, a market sentiment index is constructed. In this way, the multiple sentiment factor data within the historical time interval are screened using the obtained risk indicator data to obtain multiple valid market sentiment factor data whose correlation with the risk indicator data meets the preset conditions. This avoids the inaccurate selection problem of manually selecting sentiment data in traditional techniques. Moreover, the market sentiment index is constructed based on the valid market sentiment factor data and the risk indicator data, avoiding the inaccurate market sentiment index construction problem that exists in the unsupervised sentiment index construction method in traditional techniques, thereby improving the accuracy of market sentiment index construction.

[0069] In one embodiment, based on Figure 2 The embodiment shown, see Figure 3 , the embodiment of the present application relates to the process of constructing a market sentiment index based on risk indicator data and each effective market sentiment factor data. Figure 3 As shown, step 203 may include Figure 3 Steps 301 and 302 are shown.

[0070] Step 301 : Obtain target correlation coefficients corresponding to respective effective market sentiment factor data based on respective effective market sentiment factor data and risk indicator data.

[0071] The target correlation coefficient can represent the degree of influence of each effective market sentiment factor data on the risk indicator data. In one possible implementation, the server performs correlation analysis on each effective market sentiment factor data and the risk indicator data respectively to obtain the target correlation coefficient corresponding to each effective market sentiment factor data; in another possible implementation, the server performs correlation analysis on each effective market sentiment factor data and the risk indicator data respectively to obtain the correlation coefficient to be verified corresponding to each effective market sentiment factor data. Then, the server verifies the correlation coefficient to be verified according to certain verification conditions. If the verification is successful, the correlation coefficient to be verified is used as the target correlation coefficient. Otherwise, the server re-performs correlation analysis on each effective market sentiment factor data and the risk indicator data.

[0072] Step 302: construct a market sentiment index based on the target correlation coefficients and the effective market sentiment factor data.

[0073] In one possible implementation, the server uses each target correlation coefficient as a factor weight, and performs statistical processing on each valid market sentiment factor data according to each factor weight to obtain a market sentiment index. The statistical processing can calculate the weighted average of each valid market sentiment factor data according to each factor weight, thereby constructing each valid market sentiment factor data into a comprehensive numerical value, namely, the market sentiment index data.

[0074] In this way, in the above embodiment, the server determines the target correlation coefficient corresponding to each effective market sentiment factor data based on the screened effective market sentiment factor data and risk indicator data, thereby determining the degree of influence of each effective market sentiment factor data on the risk indicator data, and constructing market sentiment index data with higher accuracy.

[0075] In one embodiment, based on Figure 3 The embodiment shown, see Figure 4 , the embodiment of the present application relates to a process of obtaining the target correlation coefficient corresponding to each effective market sentiment factor data according to each effective market sentiment factor data and risk indicator data. Figure 4 As shown, step 301 may include Figure 4 Steps 401 to 404 are shown.

[0076] Step 401: Divide the data of each effective market sentiment factor into training factor data and testing factor data in chronological order.

[0077] In a possible implementation, the server divides the market sentiment factor data into training factor data and testing factor data in a temporal order, with the training factor data being located before the testing factor data in temporal order.

[0078] Step 402: Obtain the initial correlation coefficient corresponding to each training factor data based on each training factor data and risk indicator data.

[0079] In one possible implementation, the training factor data is used to perform correlation analysis with the risk indicator data to obtain an initial correlation coefficient to be verified. The initial correlation coefficient can characterize the degree of influence of each effective market sentiment factor data on the risk indicator data. However, in order to ensure accuracy, the server needs to verify the presented correlation coefficient before determining the final target correlation coefficient.

[0080] Step 403: construct a test sentiment index using the initial correlation coefficients and the test factor data, and verify the validity of the test sentiment index based on the risk indicator data.

[0081] The server uses each initial correlation coefficient as the initial factor weight, and then constructs the test sentiment index through each initial correlation coefficient and each test factor data. In a possible implementation, the server can verify the validity of the test sentiment index and the risk indicator data by comparison verification, such as Figure 5 As shown, the risk indicator data of the same time period is compared with the test sentiment index to determine whether the difference between the slope of the risk indicator data at the same moment and the slope of the test sentiment index at the same moment is within a preset difference threshold, or, when the test sentiment index at the same moment reaches a peak, whether the risk indicator data within the preset time range also has a peak. If the above verification conditions are met, it means that the verification is passed; otherwise, the verification fails.

[0082] Step 404: If the verification is successful, the initial correlation coefficients are used as target correlation coefficients.

[0083] In one possible implementation, when the validity verification passes, it means that the test sentiment index constructed by the server can monitor and estimate the risk indicator data. At this time, the server uses the initial correlation coefficients as the target correlation coefficients. In another possible implementation, when the validity verification fails, it means that the test sentiment index constructed by the server cannot accurately monitor and estimate the risk indicator data. At this time, the server re-performs correlation analysis on the effective market sentiment factor data and risk indicator data.

[0084] Thus, in the above embodiment, the server divides each valid factor data into training factor data and test factor data in time series, analyzes the training factor data and the risk indicator data to determine the initial correlation coefficient to be verified, and then constructs the emotion index based on the initial correlation coefficient and the test factor data to obtain the test emotion index. By effectively verifying the test emotion index, an accurate target correlation coefficient is obtained, which makes the construction of the emotion index more effective.

[0085] In one embodiment, based on Figure 2 The embodiment shown, see Figure 6 , the embodiment of the present application relates to a process of obtaining multiple valid market sentiment factor data corresponding to a historical time interval based on risk indicator data. Figure 6 As shown, step 202 may include Figure 5 Steps 601 to 603 are shown.

[0086] Step 601 : Acquire a plurality of initial market sentiment factor data corresponding to a historical time interval, and extract sentiment trend item data and sentiment detrended item data corresponding to the initial market sentiment factor data based on the initial market sentiment factor data.

[0087] The initial market sentiment factor data is the market sentiment data mentioned above. The initial market sentiment factor data is time series data. Therefore, the server can perform time series analysis on the initial market sentiment factor data to extract the sentiment trend item data and the sentiment detrended item data of the initial market sentiment factor data.

[0088] In a possible embodiment, the emotion trend item data and emotion detrended item data of the initial market emotion factor data are obtained by the server through screening based on the initial candidate emotion trend item data and the initial candidate emotion detrended item data of the initial market emotion factor data. Since it is necessary to preset an emotion moving period to extract the emotion trend item data of the initial market emotion factor data, in order to maximize the correlation between the emotion trend item data and the emotion detrended item data of the final obtained initial market emotion factor data and the risk indicator data, the server presets an emotion moving period range, for example, 20-100 days, and then, the server respectively obtains the initial candidate emotion trend item data of the initial market emotion factor data with an emotion moving period of 20 to 100 days, and the number of the initial candidate emotion trend item data is 80. The server performs a correlation analysis on each initial candidate emotion trend item data and the risk indicator data, and obtains the emotion trend item correlation coefficient corresponding to each initial candidate emotion trend item data, and selects the initial candidate emotion trend item data with the largest emotion trend item correlation coefficient as the emotion trend item data of the initial market emotion factor data.

[0089] After obtaining the sentiment trend item data of the initial market sentiment factor data, the sentiment detrended item data of the initial market sentiment factor data can be obtained by subtracting the sentiment trend item data from the initial market sentiment factor data.

[0090] Step 602 : Preprocess each emotion trend item data and each emotion detrended item data to obtain a plurality of candidate trend item data and a plurality of candidate detrended item data.

[0091] After obtaining the sentiment trend item data and sentiment detrended item data of each initial market sentiment factor data, in order to facilitate subsequent processing, the server needs to preprocess each sentiment trend item data and each sentiment detrended item data, thereby obtaining multiple candidate trend item data and multiple candidate detrended item data. The preprocessing method includes at least stationarity processing and numerical normalization processing.

[0092] In order to reduce the phenomenon of pseudo-regression of each emotion trend item data and each emotion detrended item data, it is necessary to perform stationarity processing on each emotion trend item data and each emotion detrended item data. Regarding the stationarity processing process, the server performs a stationarity test on each emotion trend item data and each emotion detrended item data. The stationarity test method can be an ADF test (Augmented Dickey-Fuller test, unit root test). When the result of the stationarity test is a success, it means that the emotion trend item data or the emotion detrended item data is stationary, and the server does not need to perform stationarity processing on the data; when the result of the stationarity test is a failure, it means that the emotion trend item data or the emotion detrended item data is not stationary, and the server needs to perform stationarity processing on the data. The method for stationarity processing can be to perform differential processing on the emotion trend item data or the emotion detrended item data. Thereby, each emotion trend item data and each emotion detrended item data are stationary data.

[0093] Since the types of initial market sentiment factor data are different, for example, when the initial market sentiment factor data is the rate of return of the financial market, the values ​​of the sentiment trend item data and the sentiment detrended item data corresponding to the initial market sentiment factor data are both between 0 and 1. However, when the initial market sentiment factor data is the number of stocks that rise and fall in the financial market, the values ​​of the sentiment trend item data and the sentiment detrended item data corresponding to the initial market sentiment factor data are both integers. In order to facilitate subsequent processing, the server needs to perform numerical normalization on each sentiment trend item data and each sentiment detrended item data. Regarding the numerical normalization process, for each sentiment trend item data or sentiment detrended item data, the server obtains the values ​​of the previous 50 days for each day, such as day t. Then, the server sorts the values ​​of these 51 days from small to large, determines the ranking of the value of day t within these 51 days, and updates the value of day t based on the ranking. For example, if the value of day t ranks 1, it means that the value of day t is the smallest value among the values ​​of these 51 days, and is updated to 1 / 51. After the daily value of each emotion trend item data or emotion detrended item data is updated according to the above method, the daily value of each emotion trend item data or emotion detrended item data is between 0 and 1.

[0094] Step 603 : Screening and processing each candidate trend item data and each candidate detrended item data according to the risk indicator data to obtain a plurality of valid market sentiment factor data.

[0095] The effective market sentiment factor data is data that can influence and preset the risk indicator data. Therefore, after the server pre-processes each sentiment trend item data and each sentiment detrended item data, it is necessary to screen and process each candidate sentiment trend item data and each candidate sentiment detrended item data.

[0096] Regarding the method for the server to screen and process each candidate sentiment trend item data and each candidate sentiment detrended item data, in one possible implementation, the server first performs single-factor screening processing on each candidate sentiment trend item data and each candidate sentiment detrended item data, and selects single-factor data that can influence and predict risk indicator data. Then, the server performs multi-factor screening processing on each single factor data, thereby obtaining a factor data combination that can influence and predict risk indicator data. The multiple factor data included in the factor data combination are effective market sentiment factor data.

[0097] In this way, in the above embodiment, the server extracts the emotion trend item data and emotion detrended item data of the initial market emotion factor data, and then pre-processes each emotion trend item data and each emotion detrended item data to obtain multiple candidate trend item data and multiple candidate detrended item data, and then screens each candidate emotion trend item data and each candidate emotion detrended item data to obtain multiple effective market emotion factor data that can influence and predict risk indicator data, and sets a screening method for the market emotion factor data for constructing the emotion index, avoiding the traditional technology of manually selecting the market emotion factor data for constructing the emotion index, and improving the accuracy of the emotion index construction.

[0098] In one embodiment, based on Figure 5 The embodiment shown, see Figure 7 The embodiment of the present application relates to a process of screening and processing each candidate trend item data and each candidate detrended item data according to risk indicator data to obtain multiple valid market sentiment factor data. Figure 7 As shown, step 503 may include Figure 7 Steps 701 and 702 are shown.

[0099] Step 701 : performing single factor screening processing on each candidate trend item data and each candidate detrended item data according to the risk indicator data to obtain a plurality of single factor data.

[0100] The purpose of the single factor screening process is to obtain multiple single factor data that can influence and predict risk indicator data. The single factor screening process includes stationarity test processing, correlation analysis processing, linear regression test processing and causal test processing.

[0101] Regarding the process of stationarity test processing, the server performs a stationarity test on each candidate emotion trend item data and each candidate emotion detrended item data. For candidate emotion trend item data or candidate emotion detrended item data that fail the stationarity test, the server will screen them out and retain the candidate emotion trend item data or candidate emotion detrended item data that pass the stationarity test.

[0102] Regarding the correlation analysis process, the server performs correlation analysis on each candidate emotion trend item data and each candidate emotion detrended item data and risk indicator data, and presets a correlation threshold. For candidate emotion trend item data or candidate emotion detrended item data whose correlation is lower than the preset correlation threshold, the server will screen them out. For candidate emotion trend item data or candidate emotion detrended item data whose correlation is not lower than the preset correlation threshold, the server will retain them.

[0103] Regarding the process of linear regression test processing, the server can perform linear regression test on each candidate emotion trend item data and each candidate emotion detrended item data and risk indicator data through t test (Student t-test). For data that fails the linear regression test, it means that there is no linear relationship between it and the risk indicator data, that is, the data cannot affect the risk indicator data, and the server will screen it out. Otherwise, it means that there is a linear relationship between the data and the risk indicator data, that is, the data can affect the risk indicator data, and the server will retain it.

[0104] Regarding causal test processing, the server can perform causal test processing on each candidate sentiment trend item data and each candidate sentiment detrended item data and risk indicator data through Granger causality test. If the test fails, it means that the data cannot predict the risk indicator data, and the server will filter it out. Otherwise, it means that the data can predict the risk indicator data, and the server will retain it.

[0105] The four processing steps of the above-mentioned single factor screening process are not in any particular order. In the end, the server obtains multiple single factor data, and each single factor data can independently affect and predict the risk indicator data.

[0106] In step 702, the server performs multi-factor screening on each single factor data according to the risk indicator data to obtain multiple valid market sentiment factor data.

[0107] Although each single factor data can independently affect and predict the risk indicator data, when multiple single factor data are combined together, they will not necessarily affect the risk indicator data. Therefore, the server needs to perform multi-factor screening processing on each single factor data, where the multi-factor screening processing includes combined linear test processing.

[0108] The combined linear test can be an F-test (F-test, joint hypothesis test). Regarding the process of combined linear test processing, the server obtains the t-value of the t-test between each single factor data and the risk index data, and then traverses each single factor data, selects the single factor data with the largest t-value and the risk index data to perform the F-test, screens out the single factor data that fails the F-test, and retains the single factor data that succeeds the F-test. Then, the server selects the single factor data with the second largest t-value from the remaining multiple single factor data, and performs the F-test together with the risk index data together with the multiple single factor data previously retained. The single factor data that fails the F-test is screened out, and the single factor data that succeeds the F-test is retained. The above method is iterated to perform the F-test on each single factor screening, and the factor data combination composed of the multiple single factor data finally retained can affect the ice and rain side risk index data. The multiple single factor data retained are used as effective market sentiment factor data.

[0109] In this way, the above embodiment provides a specific implementation method for the server to screen and process each candidate trend item data and each candidate detrended item data to obtain effective market sentiment factor data, so that there is a screening method for the effective market sentiment factor data for constructing the market sentiment index, thereby improving the accuracy of the market sentiment index construction.

[0110] In one embodiment, a method for constructing a market sentiment index is provided. The method comprises the following steps:

[0111] Step a: Obtain risk indicator data corresponding to a preset historical time interval.

[0112] Among them, the risk indicator data includes the actual resource values ​​corresponding to each unit time within the historical time interval.

[0113] Step b: obtaining a plurality of initial market sentiment factor data corresponding to a historical time interval, and extracting sentiment trend item data and sentiment detrending item data corresponding to the initial market sentiment factor data based on the initial market sentiment factor data.

[0114] Step c: pre-processing each emotion trend item data and each emotion detrended item data to obtain a plurality of candidate trend item data and a plurality of candidate detrended item data.

[0115] Step d: Based on the risk indicator data, single factor screening is performed on each candidate trend item data and each candidate detrended item data to obtain multiple single factor data.

[0116] Among them, the single factor screening process includes stationarity test processing, correlation analysis processing, linear regression test processing and causal test processing.

[0117] In step e, based on the risk indicator data, a multi-factor screening process is performed on each single factor data to obtain multiple effective market sentiment factor data. The multi-factor screening process includes a combined linear test process.

[0118] Among them, the correlation between the effective market sentiment factor data and the actual resource values ​​meets the preset conditions.

[0119] Step f: Divide the data of each effective market sentiment factor into training factor data and testing factor data in chronological order.

[0120] Step g: Obtain the initial correlation coefficient corresponding to each training factor data based on each training factor data and risk indicator data.

[0121] Step h: construct a test sentiment index using the initial correlation coefficients and the test factor data, and verify the validity of the test sentiment index based on the risk indicator data.

[0122] In step i, if the verification is successful, the initial correlation coefficients are used as the target correlation coefficients.

[0123] In step j, each target correlation coefficient is used as a factor weight, and statistical processing is performed on each effective market sentiment factor data according to each factor weight to obtain a market sentiment index.

[0124] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0125] The following, combined Figure 8 , the implementation process of the data filtering method of the embodiment of the present application is exemplarily described.

[0126] 1) The server obtains the actual resource value corresponding to each unit time in the historical time interval;

[0127] 2) The server performs logarithmic transformation on the actual resource values, extracts the detrended item data, and obtains the risk indicator data;

[0128] 3) The server obtains market sentiment data;

[0129] 4) The server pre-processes the market sentiment data, including extracting trend item data, extracting detrended item data, performing stationarity processing, and performing numerical normalization processing to obtain candidate market sentiment factor data;

[0130] 5) The server performs single-factor screening and multi-factor screening based on the risk indicator data and candidate market sentiment factor data to obtain effective market sentiment factor data.

[0131] The following, combined Figure 9 , the implementation process of the single factor screening process of the embodiment of the present application is exemplarily described.

[0132] 1) The server performs a stationarity test on the candidate market factor data and retains the market factor data that passes the stationarity test;

[0133] 2) The server performs correlation analysis on the market factor data and risk indicator data retained from the stationarity test, and retains the market factor data whose correlation with the risk indicator data is less than a correlation threshold;

[0134] 3) The server performs a linear regression test on the market factor data and risk indicator data whose correlation is less than the correlation threshold, and retains the market factor data that is linearly correlated with the risk indicator data;

[0135] 4) The server performs Granger causality test on the market factor data and risk indicator data retained in the linear regression test process to obtain single factor data that can predict the risk indicator data.

[0136] The following, combined Figure 10 , the implementation process of the multi-factor screening process of the embodiment of the present application is exemplarily described.

[0137] 1) The server traverses all single-factor data, selects the single-factor data with the largest t-value in the t-test with the risk indicator data, and puts it into the factor data combination;

[0138] 2) Perform an F test on the single factor data and risk indicator data in the single factor data combination, filter out the single factor data that are not significant in the F test, and retain the significant single factor data;

[0139] 3) Determine whether all single factor data are taken out. If so, perform F test on the last single factor data combination, filter out the single factor data that are not significant in the F test, and retain the significant multiple single factor data as the effective market sentiment factor data. Otherwise, repeat Figure 10 Steps 1) and 2) in the embodiment;

[0140] Based on the same inventive concept, embodiments of the present application also provide a market sentiment index construction device for implementing the aforementioned market sentiment index construction method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more embodiments of the market sentiment index construction device provided below can be found in the aforementioned limitations of the market sentiment index construction method and will not be further elaborated here.

[0141] In an exemplary embodiment, Figure 11 As shown, a market sentiment index construction device is provided, including: a first acquisition module 1101, a second acquisition module 1102 and a construction module 1103, wherein:

[0142] The first acquisition module 1101 is used to obtain risk indicator data corresponding to a preset historical time interval, where the risk indicator data includes actual resource values ​​corresponding to each unit time in the historical time interval;

[0143] The second acquisition module 1102 is used to obtain a plurality of effective market sentiment factor data corresponding to a historical time interval based on the risk indicator data, wherein the correlation between the effective market sentiment factor data and the actual resource value meets a preset condition;

[0144] The construction module 1103 is used to construct a market sentiment index based on the risk indicator data and the data of each effective market sentiment factor.

[0145] In one embodiment, the construction module 1103 includes:

[0146] A target correlation coefficient acquisition unit is used to acquire the target correlation coefficient corresponding to each effective market sentiment factor data based on each effective market sentiment factor data and risk indicator data;

[0147] The market sentiment index construction unit is used to construct the market sentiment index based on the target correlation coefficients and the effective market sentiment factor data.

[0148] In one embodiment, the market sentiment index construction unit is further configured to:

[0149] The target correlation coefficients are used as factor weights, and the effective market sentiment factor data are statistically processed according to the factor weights to obtain the market sentiment index.

[0150] In one embodiment, the target correlation coefficient acquisition unit is further configured to:

[0151] According to the time series, the data of each effective market sentiment factor is divided into training factor data and testing factor data;

[0152] According to each training factor data and risk indicator data, the initial correlation coefficient corresponding to each training factor data is obtained;

[0153] The test sentiment index is constructed using the initial correlation coefficients and the test factor data, and the validity of the test sentiment index is verified based on the risk indicator data;

[0154] If the verification is successful, the initial correlation coefficients are used as the target correlation coefficients.

[0155] In one embodiment, the second acquisition module 1102 includes:

[0156] An extraction unit is used to obtain a plurality of initial market sentiment factor data corresponding to a historical time interval, and extract sentiment trend item data and sentiment detrended item data corresponding to the initial market sentiment factor data based on the initial market sentiment factor data;

[0157] A preprocessing unit, configured to preprocess each emotion trend item data and each emotion detrended item data to obtain a plurality of candidate trend item data and a plurality of candidate detrended item data;

[0158] The screening unit is used to screen each candidate trend item data and each candidate detrended item data according to the risk indicator data to obtain multiple valid market sentiment factor data.

[0159] In one embodiment, the screening unit is further configured to:

[0160] Based on the risk indicator data, single factor screening is performed on each candidate trend item data and each candidate detrended item data to obtain multiple single factor data. The single factor screening process includes stationarity test processing, correlation analysis processing, linear regression test processing and causal test processing;

[0161] Based on the risk indicator data, each single factor data is subjected to multi-factor screening processing to obtain multiple effective market sentiment factor data. The multi-factor screening processing includes combined linear test processing.

[0162] Each module in the aforementioned market sentiment index construction device may be implemented in whole or in part through software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor within a computer device in the form of hardware, or may be stored in a computer device memory in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0163] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 12As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store market sentiment index construction data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a market sentiment index construction method is implemented.

[0164] Those skilled in the art will understand that Figure 12 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0165] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0166] Obtain risk indicator data corresponding to a preset historical time interval, including actual resource values ​​corresponding to each unit time within the historical time interval;

[0167] Based on the risk indicator data, multiple valid market sentiment factor data corresponding to the historical time interval are obtained, and the correlation between the valid market sentiment factor data and the actual resource value meets the preset conditions;

[0168] A market sentiment index is constructed based on risk indicator data and data of various effective market sentiment factors.

[0169] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0170] Obtain the target correlation coefficient corresponding to each effective market sentiment factor data based on each effective market sentiment factor data and risk indicator data;

[0171] The market sentiment index is constructed based on the target correlation coefficients and the effective market sentiment factor data.

[0172] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0173] The target correlation coefficients are used as factor weights, and the effective market sentiment factor data are statistically processed according to the factor weights to obtain the market sentiment index.

[0174] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0175] According to the time series, the data of each effective market sentiment factor is divided into training factor data and testing factor data;

[0176] According to each training factor data and risk indicator data, the initial correlation coefficient corresponding to each training factor data is obtained;

[0177] The test sentiment index is constructed using the initial correlation coefficients and the test factor data, and the validity of the test sentiment index is verified based on the risk indicator data;

[0178] If the verification is successful, the initial correlation coefficients are used as the target correlation coefficients.

[0179] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0180] Acquire multiple initial market sentiment factor data corresponding to a historical time interval, and extract sentiment trend item data and sentiment detrended item data corresponding to the initial market sentiment factor data based on the initial market sentiment factor data;

[0181] Preprocessing each emotion trend item data and each emotion detrended item data to obtain a plurality of candidate trend item data and a plurality of candidate detrended item data;

[0182] Based on the risk indicator data, the candidate trend item data and the candidate detrended item data are screened and processed to obtain multiple effective market sentiment factor data.

[0183] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0184] Based on the risk indicator data, single factor screening is performed on each candidate trend item data and each candidate detrended item data to obtain multiple single factor data. The single factor screening process includes stationarity test processing, correlation analysis processing, linear regression test processing and causal test processing;

[0185] Based on the risk indicator data, each single factor data is subjected to multi-factor screening processing to obtain multiple effective market sentiment factor data. The multi-factor screening processing includes combined linear test processing.

[0186] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0187] Obtain risk indicator data corresponding to a preset historical time interval, including actual resource values ​​corresponding to each unit time within the historical time interval;

[0188] Based on the risk indicator data, multiple valid market sentiment factor data corresponding to the historical time interval are obtained, and the correlation between the valid market sentiment factor data and the actual resource value meets the preset conditions;

[0189] A market sentiment index is constructed based on risk indicator data and data of various effective market sentiment factors.

[0190] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0191] Obtain the target correlation coefficient corresponding to each effective market sentiment factor data based on each effective market sentiment factor data and risk indicator data;

[0192] The market sentiment index is constructed based on the target correlation coefficients and the effective market sentiment factor data.

[0193] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0194] The target correlation coefficients are used as factor weights, and the effective market sentiment factor data are statistically processed according to the factor weights to obtain the market sentiment index.

[0195] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0196] According to the time series, the data of each effective market sentiment factor is divided into training factor data and testing factor data;

[0197] According to each training factor data and risk indicator data, the initial correlation coefficient corresponding to each training factor data is obtained;

[0198] The test sentiment index is constructed using the initial correlation coefficients and the test factor data, and the validity of the test sentiment index is verified based on the risk indicator data;

[0199] If the verification is successful, the initial correlation coefficients are used as the target correlation coefficients.

[0200] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0201] Acquire multiple initial market sentiment factor data corresponding to a historical time interval, and extract sentiment trend item data and sentiment detrended item data corresponding to the initial market sentiment factor data based on the initial market sentiment factor data;

[0202] Preprocessing each emotion trend item data and each emotion detrended item data to obtain a plurality of candidate trend item data and a plurality of candidate detrended item data;

[0203] Based on the risk indicator data, the candidate trend item data and the candidate detrended item data are screened and processed to obtain multiple effective market sentiment factor data.

[0204] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0205] Based on the risk indicator data, single factor screening is performed on each candidate trend item data and each candidate detrended item data to obtain multiple single factor data. The single factor screening process includes stationarity test processing, correlation analysis processing, linear regression test processing and causal test processing;

[0206] Based on the risk indicator data, each single factor data is subjected to multi-factor screening processing to obtain multiple effective market sentiment factor data. The multi-factor screening processing includes combined linear test processing.

[0207] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0208] Obtain risk indicator data corresponding to a preset historical time interval, including actual resource values ​​corresponding to each unit time within the historical time interval;

[0209] Based on the risk indicator data, multiple valid market sentiment factor data corresponding to the historical time interval are obtained, and the correlation between the valid market sentiment factor data and the actual resource value meets the preset conditions;

[0210] A market sentiment index is constructed based on risk indicator data and data of various effective market sentiment factors.

[0211] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0212] Obtain the target correlation coefficient corresponding to each effective market sentiment factor data based on each effective market sentiment factor data and risk indicator data;

[0213] The market sentiment index is constructed based on the target correlation coefficients and the effective market sentiment factor data.

[0214] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0215] The target correlation coefficients are used as factor weights, and the effective market sentiment factor data are statistically processed according to the factor weights to obtain the market sentiment index.

[0216] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0217] According to the time series, the data of each effective market sentiment factor is divided into training factor data and testing factor data;

[0218] According to each training factor data and risk indicator data, the initial correlation coefficient corresponding to each training factor data is obtained;

[0219] The test sentiment index is constructed using the initial correlation coefficients and the test factor data, and the validity of the test sentiment index is verified based on the risk indicator data;

[0220] If the verification is successful, the initial correlation coefficients are used as the target correlation coefficients.

[0221] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0222] Acquire multiple initial market sentiment factor data corresponding to a historical time interval, and extract sentiment trend item data and sentiment detrended item data corresponding to the initial market sentiment factor data based on the initial market sentiment factor data;

[0223] Preprocessing each emotion trend item data and each emotion detrended item data to obtain a plurality of candidate trend item data and a plurality of candidate detrended item data;

[0224] Based on the risk indicator data, the candidate trend item data and the candidate detrended item data are screened and processed to obtain multiple effective market sentiment factor data.

[0225] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0226] Based on the risk indicator data, single factor screening is performed on each candidate trend item data and each candidate detrended item data to obtain multiple single factor data. The single factor screening process includes stationarity test processing, correlation analysis processing, linear regression test processing and causal test processing;

[0227] Based on the risk indicator data, each single factor data is subjected to multi-factor screening processing to obtain multiple effective market sentiment factor data. The multi-factor screening processing includes combined linear test processing.

[0228] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0229] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0230] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0231] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for constructing a market sentiment index, characterized in that: The method comprises: Obtain risk indicator data corresponding to a preset historical time interval, wherein the risk indicator data includes actual resource values ​​corresponding to each unit time in the historical time interval; According to the risk indicator data, a plurality of valid market sentiment factor data corresponding to a historical time interval is obtained, and the correlation between the valid market sentiment factor data and the actual resource value meets a preset condition; A market sentiment index is constructed based on the risk indicator data and the effective market sentiment factor data.

2. The method according to claim 1, characterized in that The step of constructing a market sentiment index according to the risk indicator data and each of the effective market sentiment factor data comprises: According to each of the effective market sentiment factor data and the risk indicator data, respectively, a target correlation coefficient corresponding to each of the effective market sentiment factor data is obtained; The market sentiment index is constructed according to each of the target correlation coefficients and each of the effective market sentiment factor data.

3. The method according to claim 2, characterized in that The constructing the market sentiment index according to each of the target correlation coefficients and each of the effective market sentiment factor data comprises: Each of the target correlation coefficients is used as a factor weight, and statistical processing is performed on each of the effective market sentiment factor data according to each of the factor weights to obtain the market sentiment index.

4. The method according to claim 2, characterized in that: The step of obtaining the target correlation coefficient corresponding to each of the effective market sentiment factor data and the risk indicator data respectively includes: According to the time series order, the data of each effective market sentiment factor is divided into training factor data and testing factor data; Obtaining initial correlation coefficients corresponding to the training factor data respectively according to the training factor data and the risk indicator data; Constructing a test sentiment index using each of the initial correlation coefficients and each of the test factor data, and verifying the validity of the test sentiment index based on the risk indicator data; If the verification is successful, each of the initial correlation coefficients is used as each of the target correlation coefficients.

5. The method according to claim 1, characterized in that The step of obtaining a plurality of effective market sentiment factor data corresponding to a historical time interval according to the risk indicator data includes: Acquire multiple initial market sentiment factor data corresponding to the historical time interval, and extract sentiment trend item data and sentiment detrending item data corresponding to the initial market sentiment factor data according to the initial market sentiment factor data; Preprocessing each of the emotion trend item data and each of the emotion detrended item data to obtain a plurality of candidate trend item data and a plurality of candidate detrended item data; According to the risk indicator data, each of the candidate trend item data and each of the candidate detrended item data is screened and processed to obtain a plurality of the effective market sentiment factor data.

6. The method according to claim 5, characterized in that The step of screening the candidate trend item data and the candidate detrended item data according to the risk indicator data to obtain a plurality of the effective market sentiment factor data includes: According to the risk indicator data, a single factor screening process is performed on each of the candidate trend item data and each of the candidate detrended item data to obtain a plurality of single factor data, wherein the single factor screening process includes a stationarity test process, a correlation analysis process, a linear regression test process and a causal test process; According to the risk indicator data, a multi-factor screening process is performed on each of the single factor data to obtain a plurality of the effective market sentiment factor data, and the multi-factor screening process includes a combined linear test process.

7. A market sentiment index construction device, characterized in that: The device comprises: A first acquisition module is used to acquire risk indicator data corresponding to a preset historical time interval, wherein the risk indicator data includes actual resource values ​​corresponding to each unit time in the historical time interval; A second acquisition module is used to acquire a plurality of effective market sentiment factor data corresponding to a historical time interval according to the risk indicator data, wherein the correlation between the effective market sentiment factor data and the actual resource value meets a preset condition; A construction module is used to construct a market sentiment index based on the risk indicator data and the effective market sentiment factor data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images