Risk prediction method and device and electronic equipment

Abstract

The invention discloses a risk prediction method and device and electronic equipment. The method relates to the field of artificial intelligence, and comprises the steps that a preset risk variable is determined, a financial feature set of a target user is acquired from a block chain database, the financial feature set comprises M financial features and feature values of the financial features, and M is a positive integer; according to the correlation degree and the gain value between the preset risk variable and the financial feature set, financial features associated with the preset risk variable are selected from the financial feature set, N target features are obtained, and N is a positive integer; and inputting the feature values of the N target features into a risk prediction model associated with a preset risk variable to obtain a financial risk of the target user under a preset risk variable dimension. According to the method and the device, the problem of low accuracy of determining financial risks through artificial experience in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and specifically, to a risk prediction method, device, and electronic device. Background Art

[0002] With the increasing complexity of the financial market and the rapid development of big data technology, the financial risks associated with financial services are also increasing. In order to ensure the safety of users' assets, financial institutions need to predict financial risks and provide timely feedback to users on the prediction results, so as to give users early warning.

[0003] However, currently, when predicting financial risks, people often rely on the experience and judgment of business personnel. For example, business personnel judge whether there are certain financial risks based on the user's transaction situation, which makes it difficult to make comprehensive and accurate risk predictions.

[0004] Currently, no effective solution has been proposed to address the problem of low accuracy in determining financial risks through manual experience in related technologies. Summary of the invention

[0005] The main purpose of this application is to provide a risk prediction method, device and electronic device to solve the problem of low accuracy in determining financial risks through manual experience in related technologies.

[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a risk prediction method is provided. The method includes: determining a preset risk variable, and obtaining a financial feature set of a target user from a blockchain database, wherein the financial feature set includes M financial features and feature values ​​of each financial feature, and M is a positive integer; selecting financial features associated with the preset risk variable from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set, and obtaining N target features, wherein N is a positive integer; inputting the feature values ​​of the N target features into a risk prediction model associated with the preset risk variable, and obtaining the financial risk of the target user under the preset risk variable dimension.

[0007] Optionally, selecting financial features associated with preset risk variables from the financial feature set according to the correlation and gain values ​​between the preset risk variables and the financial feature set to obtain N target features includes: obtaining the correlation and gain values ​​between each financial feature in the financial feature set and the preset risk variables to obtain M groups of parameter sets, wherein each group of parameter sets corresponds to the correlation and gain value of a financial feature; obtaining weights of the correlation and gain values, and performing weighted summation of the parameters in each group of parameter sets according to the weights to obtain M scores; obtaining the number of parameters required for using the risk prediction model, and selecting financial features from the M financial features in descending order of scores according to the number of parameters to obtain N target features.

[0008] Optionally, obtaining the correlation between each financial feature in the financial feature set and the preset risk variable includes: for any financial feature, obtaining historical feature values ​​of the financial feature and historical variable values ​​of the preset risk variable at multiple historical moments to obtain multiple sets of target numerical values; calculating a first average value of the financial feature and a second average value of the preset risk variable based on the multiple sets of target numerical values; calculating the covariance between the financial feature and the preset risk variable based on the multiple sets of target numerical values, the first average value and the second average value; calculating a first standard deviation of the financial feature and a second standard deviation of the preset risk variable based on the multiple sets of target numerical values, the first average value and the second average value; and determining the correlation based on the first standard deviation, the second standard deviation and the covariance.

[0009] Optionally, obtaining the gain value between each financial feature in the financial feature set and the preset risk variable includes: for any financial feature, obtaining the historical feature value of the financial feature and the historical variable value of the preset risk variable at multiple historical moments to obtain multiple sets of target numerical values; calculating the entropy of the preset risk variable and the conditional entropy of the preset risk variable under the financial feature as the condition based on the multiple sets of target numerical values; determining the gain value based on the entropy of the preset risk variable and the conditional entropy.

[0010] Optionally, obtaining historical characteristic values ​​of financial characteristics and historical variable values ​​of preset risk variables at multiple historical moments to obtain multiple sets of target numerical values ​​includes: inputting the account information and private key of the target user into the blockchain database; when the private key is correct, receiving multiple sets of initial data sets of the target user sent by the blockchain database; judging whether there are candidate data sets with missing data in the multiple sets of initial data sets; when there are candidate data sets, calculating the missing values ​​based on the financial characteristic values ​​contained in the candidate data sets, and filling the missing values ​​into the multiple sets of initial data sets to obtain multiple sets of target numerical values.

[0011] Optionally, calculating the missing value based on the financial characteristic values ​​included in the candidate data set includes: obtaining the position sequence number of the missing value in the candidate data set, and obtaining multiple first characteristic values ​​with the smallest difference with the position sequence number; calculating the adjustment coefficient of the missing value based on the financial characteristic values ​​included in the candidate data set; and calculating the missing value based on the adjustment coefficient and the multiple first characteristic values.

[0012] Optionally, before inputting the characteristic values ​​of N target characteristics into the risk prediction model associated with the preset risk variables, the method also includes: obtaining historical characteristic values ​​of the N target characteristics to obtain multiple groups of historical characteristic values; obtaining the risk coefficient of the preset risk variable corresponding to each group of historical characteristic values ​​to obtain multiple risk coefficients; determining each group of historical characteristic values ​​and the corresponding risk coefficient as a group of sample data to obtain multiple groups of sample data; training the neural network model through the multiple groups of sample data to obtain a risk prediction model.

[0013] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a risk prediction device is provided. The device includes: a first acquisition unit, which is used to determine a preset risk variable and obtain a financial feature set of a target user from a blockchain database, wherein the financial feature set includes M financial features and feature values ​​of each financial feature, and M is a positive integer; a selection unit, which is used to select financial features associated with the preset risk variable from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set, and obtain N target features, wherein N is a positive integer; a prediction unit, which is used to input the feature values ​​of the N target features into a risk prediction model associated with the preset risk variable, and obtain the financial risk of the target user under the preset risk variable dimension.

[0014] Optionally, the selection unit includes: a first acquisition subunit, used to obtain the correlation and gain value between each financial feature in the financial feature set and the preset risk variable, and obtain M groups of parameter sets, wherein each group of parameter sets corresponds to the correlation and gain value of a financial feature; a second acquisition subunit, used to obtain the weights of the correlation and gain values, and perform weighted summation of the parameters in each group of parameter sets according to the weights to obtain M scores; a third acquisition subunit, used to obtain the number of parameters required for using the risk prediction model, and select financial features from the M financial features in descending order of scores according to the number of parameters to obtain N target features.

[0015] Optionally, the first acquisition subunit includes: a first acquisition module, used to obtain, for any financial feature, historical feature values ​​of the financial feature and historical variable values ​​of preset risk variables at multiple historical moments to obtain multiple sets of target numerical values; a first calculation module, used to calculate a first average value of the financial feature and a second average value of the preset risk variable based on multiple sets of target numerical values; a second calculation module, used to calculate the covariance between the financial feature and the preset risk variable based on multiple sets of target numerical values, the first average value and the second average value; a third calculation module, used to calculate the first standard deviation of the financial feature and the second standard deviation of the preset risk variable based on multiple sets of target numerical values, the first average value and the second average value; a first determination module, used to determine the correlation based on the first standard deviation, the second standard deviation and the covariance.

[0016] Optionally, the first acquisition subunit includes: a second acquisition module, used to obtain, for any financial feature, historical feature values ​​of the financial feature at multiple historical moments and historical variable values ​​of a preset risk variable to obtain multiple sets of target numerical values; a fourth calculation module, used to calculate the entropy of the preset risk variable and the conditional entropy of the preset risk variable under the financial feature as a condition based on the multiple sets of target numerical values; a second determination module, used to determine the gain value based on the entropy and conditional entropy of the preset risk variable.

[0017] Optionally, the first acquisition module includes: an input submodule, used to input the account information and private key of the target user into the blockchain database; a receiving submodule, used to receive multiple groups of initial data sets of the target user sent by the blockchain database when the private key is correct; a judgment submodule, used to judge whether there are candidate data sets with missing data in the multiple groups of initial data sets: a first calculation submodule, used to calculate the missing values ​​according to the financial feature values ​​contained in the candidate data sets when there are candidate data sets, and fill the missing values ​​into the multiple groups of initial data sets to obtain multiple groups of target numerical value sets.

[0018] Optionally, the first calculation submodule includes: an acquisition submodule, used to obtain the position sequence number of the missing value in the candidate data set, and obtain multiple first eigenvalues ​​with the smallest difference from the position sequence number; a second calculation submodule, used to calculate the adjustment coefficient of the missing value based on the financial eigenvalue contained in the candidate data set; and a third calculation submodule, used to calculate the missing value based on the adjustment coefficient and multiple first eigenvalues.

[0019] Optionally, before inputting the characteristic values ​​of N target characteristics into the risk prediction model associated with the preset risk variables, the device also includes: a second acquisition unit, used to acquire historical characteristic values ​​of the N target characteristics to obtain multiple groups of historical characteristic values; a third acquisition unit, used to acquire the risk coefficient of the preset risk variable corresponding to each group of historical characteristic values ​​to obtain multiple risk coefficients; a determination unit, used to determine each group of historical characteristic values ​​and the corresponding risk coefficient as a group of sample data to obtain multiple groups of sample data; and a training unit, used to train the neural network model through multiple groups of sample data to obtain a risk prediction model.

[0020] In order to achieve the above-mentioned purpose, according to another aspect of the present application, an electronic device is provided, the electronic device comprising a memory storing an executable program; and a processor for running the program, wherein the above-mentioned risk prediction method is executed when the program is running.

[0021] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a computer program product is provided, including computer instructions, which implement the steps of the above-mentioned risk prediction method when executed by a processor.

[0022] In an embodiment of the present application, a preset risk variable is determined, and a financial feature set of a target user is obtained from a blockchain database, wherein the financial feature set includes M financial features and feature values ​​of each financial feature, and M is a positive integer; a financial feature associated with the preset risk variable is selected from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set to obtain N target features, wherein N is a positive integer; the feature values ​​of the N target features are input into a risk prediction model associated with the preset risk variable to obtain a method for the financial risk of the target user under the preset risk variable dimension, by obtaining the financial feature set of the user, and selecting from the financial feature set a financial feature that is associated with the preset risk variable to be predicted, thereby obtaining the target feature, and then using the target feature to predict the financial risk corresponding to the preset risk variable, thereby achieving the purpose of selecting features according to different financial risks and accurately predicting the risk according to the features, thereby achieving the technical effect of improving the accuracy of financial risk prediction, and then solving the technical problem of low accuracy in determining financial risks through artificial experience in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0024] Figure 1 A hardware structure block diagram of a computer terminal for implementing a risk prediction method is shown;

[0025] Figure 2 is a flow chart of the risk prediction method provided in Example 1 of the present application;

[0026] Figure 3 is a schematic diagram of a risk prediction device provided according to Example 2 of the present application;

[0027] Figure 4 It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] It should be noted that the risk prediction method, device and electronic device determined in the present disclosure can be used in the field of artificial intelligence, and can also be used in any field other than the field of artificial intelligence. The application field of the risk prediction method, device and electronic device determined in the present disclosure is not limited.

[0032] It should be noted that the collected information, 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.) used in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with the relevant laws, regulations and standards of the relevant regions, necessary confidentiality measures are taken, and public order and good customs are not violated. Corresponding operation entrances are provided for users to choose to authorize use or refuse use. If the user chooses to refuse, the expert decision-making process is entered. For example, an interface is set between this system and relevant users or institutions. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or institution through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or institution.

[0033] The embodiments or examples of the present disclosure are not exhaustive, but are only illustrative of some embodiments or examples, and are not intended to be specific limitations on the scope of protection of the present disclosure. In the absence of contradiction, each step in a certain embodiment or example can be implemented as an independent example, and the steps can be combined arbitrarily. For example, the scheme after removing some steps in a certain embodiment or example can also be implemented as an independent example, and the order of the steps in a certain embodiment or example can be arbitrarily exchanged. In addition, the optional methods or optional examples in a certain embodiment or example can be combined arbitrarily; in addition, the various embodiments or examples can be combined arbitrarily, for example, some or all steps of different embodiments or examples can be combined arbitrarily, and a certain embodiment or example can be combined arbitrarily with the optional methods or optional examples of other embodiments or examples.

[0034] For the convenience of description, some nouns or terms involved in the embodiments of the present application are explained below:

[0035] Pre-set risk variables: refers to the pre-defined risk categories or indicators in financial risk forecasting, such as credit default risk, market volatility risk, etc., which are the main focus of risk assessment.

[0036] Blockchain database: A distributed database that uses blockchain technology to store data. It has the characteristics of decentralization, immutability and high transparency, providing guarantees for the secure storage and management of financial data.

[0037] Example 1

[0038] According to an embodiment of the present application, an embodiment of a risk prediction method is also provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0039] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal for implementing a risk prediction method. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more (102a, 102b, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0040] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0041] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the risk prediction method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, to implement the above-mentioned risk prediction method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0042] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0043] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0044] Under the above operating environment, this application provides Figure 2 The risk prediction method shown. Figure 2 is a flow chart of the risk prediction method provided in Example 1 of the present application, such as Figure 2 As shown, the method includes:

[0045] Step S201, determine the preset risk variables, and obtain the financial feature set of the target user from the blockchain database, wherein the financial feature set includes M financial features and the feature value of each financial feature, and M is a positive integer.

[0046] It should be noted that preset risk variables refer to pre-defined risk categories or indicators in financial risk forecasting, such as credit default risk, market volatility risk, etc., which are the main focus of risk assessment. For different risk variables, different financial characteristic data need to be obtained, and then the financial risks can be accurately predicted based on different financial characteristic data.

[0047] It should be noted that the financial feature set may include a set of M financial features and their corresponding feature values. These features may be transaction frequency, capital flow, etc., and are important bases for assessing financial risks.

[0048] Specifically, after determining the preset risk variables, it is necessary to obtain a set of financial features of the target user from the database based on the preset risk variables, thereby obtaining relevant financial data of the user for whom risk prediction is required, and then predicting the probability of the user encountering financial risks under the preset risk variable dimensions based on the user's financial data, wherein the user's financial data may be the financial features of the target user and the feature values ​​corresponding to the financial features. For example, the financial features may be transaction frequency, and the feature value may be 5 times / day.

[0049] For example, in order to determine whether the user has an account deposit risk, the user's financial data such as deposit frequency, withdrawal frequency, and transfer frequency may be obtained to ensure that it is possible to accurately determine whether the user has an account deposit risk.

[0050] It should be noted that a blockchain database can be used to store users' financial data, thereby ensuring the security of users' financial data.

[0051] This step first clarifies the goal of risk prediction, that is, the preset risk variables, and then extracts all relevant financial data of the target user from the blockchain database to form a financial feature set. By obtaining the user's financial feature set, it can ensure that subsequent analysis is based on comprehensive and reliable data, providing a data basis for subsequent financial risk prediction, while the use of the blockchain database ensures the authenticity and integrity of the data and reduces the possibility of data tampering.

[0052] Step S202: selecting financial features associated with the preset risk variables from the financial feature set according to the correlation and gain value between the preset risk variables and the financial feature set, to obtain N target features, where N is a positive integer.

[0053] It should be noted that the gain value may be the information gain value between the preset risk variable and the financial feature, so as to evaluate the contribution of the feature to the prediction target.

[0054] Specifically, after obtaining the financial feature set, since different characteristic values ​​of financial features are needed to perform risk prediction for different preset risk variables, in order to ensure the accuracy of risk prediction, it is necessary to first determine the similarity and gain value between the preset risk variable and each financial feature in the financial feature set, and then select financial features with higher correlation with the preset risk variable based on the similarity and gain value, that is, identify which features are most critical in risk prediction, that is, features that have a significant impact on the risk variable, thereby obtaining N target features, and then predicting the risk of the preset risk variable based on multiple target features, thereby ensuring the accuracy of the prediction results.

[0055] It should be noted that when constructing a prediction model, the historical feature values ​​of the selected target features can also be used as training samples to construct a risk prediction model, thereby reducing the complexity of the model, avoiding the interference of irrelevant features, and improving the accuracy and efficiency of the prediction model.

[0056] Step S203, inputting the characteristic values ​​of the N target characteristics into the risk prediction model associated with the preset risk variable to obtain the financial risk of the target user under the preset risk variable dimension.

[0057] It should be noted that the risk prediction model can be a model built based on a machine learning algorithm (such as a deep learning network) to predict the possibility or degree of a specific risk variable.

[0058] Specifically, after obtaining N target features, the feature values ​​of M target features can be input into the risk prediction model, such as into the long short-term memory network (LSTM), so that the model can output the target user's financial risk assessment under the preset risk variable dimension, for example, predicting the possibility of user default in the future or the impact of market fluctuations on user assets. In this way, the selected financial features are converted into financial risk assessments with practical guiding significance, providing decision support for users and helping them better manage risks.

[0059] The risk prediction method provided in the embodiment of the present application determines a preset risk variable and obtains a financial feature set of a target user from a blockchain database, wherein the financial feature set includes M financial features and feature values ​​of each financial feature, and M is a positive integer; selects financial features associated with the preset risk variable from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set, and obtains N target features, wherein N is a positive integer; inputs the feature values ​​of the N target features into a risk prediction model associated with the preset risk variable to obtain the financial risk of the target user under the preset risk variable dimension, obtains the financial feature set of the user, and selects financial features associated with the preset risk variable to be predicted from the financial feature set, thereby obtaining the target feature, and then using the target feature to predict the financial risk corresponding to the preset risk variable, thereby achieving the purpose of selecting features according to different financial risks and accurately predicting risks according to the features, thereby achieving the technical effect of improving the accuracy of financial risk prediction, and then solving the technical problem of low accuracy in determining financial risks through artificial experience in related technologies.

[0060] Optionally, in the risk prediction method provided in the embodiment of the present application, financial features associated with the preset risk variables are selected from the financial feature set according to the correlation and gain values ​​between the preset risk variables and the financial feature set to obtain N target features, including: obtaining the correlation and gain values ​​between each financial feature in the financial feature set and the preset risk variables to obtain M groups of parameter sets, wherein each group of parameter sets corresponds to the correlation and gain value of a financial feature; obtaining weights of the correlation and gain values, and performing weighted summation of the parameters in each group of parameter sets according to the weights to obtain M scores; obtaining the number of parameters required for using the risk prediction model, and selecting financial features from the M financial features in descending order of scores according to the number of parameters to obtain N target features.

[0061] It should be noted that the correlation can be used as a statistical indicator to measure the strength of the linear relationship between two variables, and the gain value is used to evaluate the contribution of the feature to the prediction of the target variable. When determining the financial features related to the preset risk variables, the correlation and gain value can be comprehensively analyzed to accurately obtain the financial features with a high correlation with the preset risk variables.

[0062] Specifically, we first analyze the relationship between each financial feature in the financial feature set and the preset risk variable, calculate the correlation and gain value between them, and form M sets of parameters, each of which corresponds to a financial feature. This process involves in-depth mining of historical data, including but not limited to fluctuations in account balances, transaction frequency in a specific time period, dynamics of capital flows, etc., and quantification of the relationship between the corresponding preset risk variables (such as default risk). By calculating the correlation and gain value, we can preliminarily identify which financial features have the strongest correlation with the preset risk variables, provide a quantitative basis for subsequent feature screening, and ensure that the model can make risk predictions based on the most valuable information.

[0063] After obtaining the gain value and correlation, it is necessary to set the weights of the correlation and gain value, and perform weighted summation of the correlation and gain values ​​in the M sets of parameters based on these two weights to obtain M scores. Then, the correlation between each financial feature and the preset risk variable is quantified based on the score to ensure that the score can fully reflect the correlation strength between the feature and the risk prediction, thereby improving the accuracy of the risk prediction model.

[0064] It should be noted that since different models have different numbers of input parameters, in order to avoid overfitting problems caused by too many irrelevant or redundant features, financial features with higher correlation can be selected in descending order of scores based on the number of input parameters required by the model, thereby obtaining N financial features with the highest correlation with the preset risk variables. In this way, while obtaining more accurate financial features, it is ensured that the model can focus on those financial features that have the greatest influence on the prediction of the preset risk variables, thereby improving the efficiency and accuracy of the prediction.

[0065] This embodiment achieves effective screening of financial features by quantitatively evaluating the relationship between financial features and preset risk variables. First, by calculating the correlation and gain value, the contribution of the features to risk prediction is quantified; second, by setting the weights of the correlation and gain value, the screening criteria can be flexibly adjusted according to specific prediction needs; finally, by sorting the scores and selecting the features with the highest N scores as the target features, it is ensured that the construction of the risk prediction model is based on the most relevant and most predictive feature set. Then, risk prediction can be performed based on the selected target features, and the risk prediction model can be trained using the historical feature values ​​of the selected target features, thereby effectively reducing the computational burden in the model training and prediction process on the basis of ensuring the accuracy of risk prediction, avoiding overfitting, and significantly enhancing the generalization ability and accuracy of the prediction model, providing financial institutions with a more accurate and efficient risk prediction method.

[0066] Optionally, in the risk prediction method provided in the embodiment of the present application, obtaining the correlation between each financial feature in the financial feature set and the preset risk variable includes: for any financial feature, obtaining the historical feature value of the financial feature and the historical variable value of the preset risk variable at multiple historical moments to obtain multiple sets of target numerical values; calculating the first average value of the financial feature and the second average value of the preset risk variable based on the multiple sets of target numerical values; calculating the covariance between the financial feature and the preset risk variable based on the multiple sets of target numerical values, the first average value and the second average value; calculating the first standard deviation of the financial feature and the second standard deviation of the preset risk variable based on the multiple sets of target numerical values, the first average value and the second average value; determining the correlation based on the first standard deviation, the second standard deviation and the covariance.

[0067] Specifically, when calculating the correlation between each financial feature and the preset risk variable, it is first necessary to obtain the historical feature values ​​of any financial feature at multiple historical moments, as well as the historical variable values ​​of the preset risk variable at each historical moment. By collecting historical data, a basis for subsequent statistical analysis is provided to ensure that the correlation calculation is based on actual financial activity data, thereby improving the reliability and accuracy of the prediction model.

[0068] After obtaining multiple sets of target numerical values, the first average value of each financial feature can be calculated based on the multiple values ​​of each feature in the multiple sets of target numerical values, and the second average value of the preset risk variable can be calculated, thereby providing the necessary data points for the subsequent calculation of covariance and standard deviation, and further helping to determine the strength of the correlation between the feature and the risk.

[0069] After obtaining the average value, the covariance between the financial feature and the preset risk variable can be calculated based on the average value, as well as the first standard deviation of the financial feature and the second standard deviation of the preset risk variable. Then, based on the combination of the covariance, the first standard deviation of the financial feature and the second standard deviation of the preset risk variable, the correlation between the financial feature and the preset risk variable can be determined using the correlation calculation formula. Thus, the linear relationship strength between the financial feature and the risk variable is accurately quantified through the correlation, providing a clear quantitative indicator for the subsequent selection of financial features that are highly correlated with risk.

[0070] For example, the correlation can be calculated through the following process: where the financial feature is the rate of return and the preset risk variable is volatility:

[0071] Table 1 shows the historical characteristic values ​​of the rate of return and the historical variable values ​​of volatility.

[0072] Financial characteristics (yield) Risk variable (volatility) 5 2 3 1.5 7 2.5 4 2 6 3

[0073] First, based on the data in Table 1, the first average value of the financial characteristics and the second average value of the preset risk variables are calculated respectively:

[0074] The first average value of financial characteristics: μX = 55 + 3 + 7 + 4 + 6 = 5;

[0075] The second average value of the preset risk variable: μY=52+1.5+2.5+2+3=2.2;

[0076] After obtaining the first average value and the second average value, the covariance Cov(X,Y) between the financial feature and the preset risk variable can be calculated by the covariance formula, and the first standard deviation σX and the second standard deviation σY can be calculated by the standard deviation formula. The correlation between the financial feature and the preset risk variable can be calculated by the following formula:

[0077] R=Cov(X,Y) / (σX·σY)=0.68;

[0078] This completes the calculation operation of the correlation between the financial characteristics and the preset risk variables.

[0079] This embodiment calculates the correlation by historical data, and realizes the accurate quantification of the correlation between financial features and preset risk variables. This method ensures the data-driven feature screening process and improves the accuracy and effectiveness of the risk prediction model.

[0080] Optionally, in the risk prediction method provided in the embodiment of the present application, obtaining the gain value between each financial feature in the financial feature set and the preset risk variable includes: for any financial feature, obtaining the historical feature value of the financial feature and the historical variable value of the preset risk variable at multiple historical moments to obtain multiple sets of target numerical values; calculating the entropy of the preset risk variable and the conditional entropy of the preset risk variable under the financial feature as the condition based on the multiple sets of target numerical values; determining the gain value based on the entropy of the preset risk variable and the conditional entropy.

[0081] Specifically, when calculating the gain value between each financial feature and the preset risk variable, it is first necessary to obtain the historical feature value of any financial feature at multiple historical moments, as well as the historical variable value of the preset risk variable at each historical moment. By collecting historical data, a basis is provided for subsequent statistical analysis, ensuring that the gain value calculation is based on actual financial activity data, thereby improving the reliability and accuracy of the prediction model.

[0082] After obtaining multiple sets of target numerical values, we can calculate the entropy H(Y) of the preset risk variable and the conditional entropy H(Y|X) of the preset risk variable under the financial feature as a condition, and then calculate the gain value of the financial feature based on the entropy and conditional entropy to obtain more accurate gain value information.

[0083] For example, the gain value can be calculated by the following process:

[0084] The preset risk variable can be "whether to default" (the value is 0 for no default, and 1 for default), and the financial feature can be "credit score". First, multiple sets of target value sets need to be obtained, as shown in Table 2.

[0085] Table 2

[0086] Serial number Credit score Whether it is a breach of contract 1 750 0 2 680 1 3 720 0 4 550 1 5 800 0

[0087] Furthermore, we need to calculate the entropy H(Y) of the entire data set:

[0088] Calculate the probability distribution of the target variable "whether to default": P(Y=0)=number of non-default samples / total number of samples, P(Y=1)=number of default samples / total number of samples, and use the entropy formula to calculate H(Y).

[0089] Furthermore, the conditional entropy H(Y|X) of the feature "credit score" is calculated:

[0090] First, split the data set into different intervals of "credit score" (or specific values ​​if the data set is small). For each subset within each interval, calculate the entropy of the target variable "whether to default". Use the formula of conditional entropy and combine the probability of each interval to calculate H(Y|X).

[0091] Finally, the information gain value IG(X)=H(Y)-H(Y|X) is calculated, thereby completing the calculation of the gain value.

[0092] For example:

[0093] The “credit score” is divided into two intervals: low score (<700) and high score (≥700), and the corresponding probability and entropy are calculated.

[0094] Low score interval: P(X=low score)=0.4, P(Y=0|X=low score)=0.3, P(Y=1|X=low score)=0.7;

[0095] High score interval: P(X=high score)=0.6, P(Y=0|X=high score)=0.8, P(Y=1|X=high score)=0.2;

[0096] Calculate H(Y|X):

[0097] H(Y|X)=P(X=low score)H(Y|X=low score)+P(X=high score)H(Y|X=high score);

[0098] Among them, H(Y|X=low score) and H(Y|X=high score) are calculated using the entropy formula respectively.

[0099] Calculate IG(X)=H(Y)-H(Y|X).

[0100] Thus, the calculation of the gain value is completed through the above calculation process.

[0101] This embodiment accurately calculates the gain value through historical data, and further improves the accuracy of the information value of the evaluation feature in risk prediction through the gain value, thereby ensuring the accuracy of the subsequent selection of target features from the financial features.

[0102] Optionally, in the risk prediction method provided in the embodiment of the present application, historical characteristic values ​​of financial characteristics and historical variable values ​​of preset risk variables at multiple historical moments are obtained to obtain multiple sets of target numerical values, including: inputting the account information and private key of the target user into the blockchain database; when the private key is correct, receiving multiple sets of initial data sets of the target user sent by the blockchain database; judging whether there are candidate data sets with missing data in the multiple sets of initial data sets; when there are candidate data sets, calculating the missing values ​​based on the financial characteristic values ​​contained in the candidate data sets, and filling the missing values ​​into the multiple sets of initial data sets to obtain multiple sets of target numerical values.

[0103] Specifically, when obtaining a target value set, it is first necessary to ensure the secure acquisition of the data, while also ensuring the accuracy and completeness of the data.

[0104] In terms of ensuring the secure acquisition of data, when acquiring data, you first need to use the target user's account information and private key to initiate a data request to the blockchain database, authenticate your identity through the private key, ensure the legitimacy of the data request, and then safely retrieve the target user's historical data from the blockchain database, laying a secure foundation for subsequent data processing and analysis.

[0105] Optionally, after verifying that the private key is correct, the blockchain database sends multiple sets of initial data containing historical data of the target user to the requesting party, thereby ensuring that the data obtained from the blockchain database is complete and accurate.

[0106] After obtaining multiple sets of initial data sets, in order to ensure the integrity of the data when processing the data subsequently, it is necessary to check the data in the multiple sets of initial data sets to determine whether there is missing data. If there is missing data, these data sets are marked as candidate data sets that need to be processed, and the missing values ​​are calculated based on the financial characteristic values ​​contained in the candidate data sets, and the missing values ​​are filled into the multiple sets of initial data sets to obtain multiple sets of target value sets. In this way, in the data preprocessing stage, the missing situations in the data sets can be discovered in time, and the missing values ​​can be generated and filled based on the existing data, thereby ensuring the integrity and consistency of all data sets in the multiple sets of target value sets, so that the financial characteristic values ​​and preset risk variable values ​​at each historical moment can be used in subsequent calculation steps, thereby improving the accuracy and reliability of subsequent calculation correlation.

[0107] This embodiment obtains data from the blockchain and generates and fills in missing data, thereby ensuring the integrity and availability of the data set, avoiding errors in subsequent calculation and prediction processes due to missing data, and thus improving the accuracy of risk prediction.

[0108] Optionally, in the risk prediction method provided in the embodiment of the present application, calculating the missing value based on the financial characteristic value contained in the candidate data set includes: obtaining the position sequence number of the missing value in the candidate data set, and obtaining multiple first characteristic values ​​with the smallest difference with the position sequence number; calculating the adjustment coefficient of the missing value based on the financial characteristic value contained in the candidate data set; calculating the missing value based on the adjustment coefficient and multiple first characteristic values.

[0109] Specifically, when calculating the missing value, all data under the feature can be sorted in chronological order to obtain a feature value sequence, and the position of the missing value in the sequence can be determined.

[0110] After the position is determined, the two first eigenvalues ​​closest to the position number of the position can be obtained. For example, when the position number (row number) of the missing value is 3, the first eigenvalues ​​obtained are the eigenvalues ​​with numbers 2 and 4, that is, the eigenvalues ​​with row numbers 2 and 4.

[0111] Furthermore, it is necessary to calculate the adjustment coefficient and calculate the eigenvalue of the missing value through the adjustment coefficient and the first eigenvalue, thereby completing the missing value generation operation.

[0112] For example, the adjustment coefficient can be calculated by the following formula:

[0113]

[0114] Among them, f i is the adjustment coefficient, j is the row number, V jis the value of the jth row, i is the position number of the missing value, that is, the row number, and I is the total number of rows.

[0115] When 1<i<I, the missing value can be calculated according to the adjustment coefficient and the first eigenvalue by the following formula:

[0116]

[0117] Among them, V i is a missing value, a is the row number that is closest to the missing value and is less than i, b is the row number that is closest to the missing value and is greater than i, V a is the first eigenvalue of the ath row, V b is the first eigenvalue of the b-th row.

[0118] When 1=i or i=I, the missing value can be calculated according to the adjustment coefficient and the first eigenvalue by the following formula:

[0119]

[0120] Where i=1, a is the row number closest to the missing value, b is the row number second closest to the missing value, and V a is the first eigenvalue of the ath row, V b is the first eigenvalue of the bth row. In the case of i=I, b is the row number closest to the missing value, a is the row number second closest to the missing value, V a is the first eigenvalue of the ath row, V b is the first eigenvalue of the b-th row.

[0121] This embodiment calculates the missing value by using other characteristic values ​​that are the same as the characteristics of the missing value, thereby ensuring the integrity of the characteristic values ​​of each characteristic obtained, and further ensuring the accuracy of the subsequent calculation of the correlation using the characteristic value.

[0122] Optionally, in the risk prediction method provided in the embodiment of the present application, before inputting the characteristic values ​​of N target characteristics into the risk prediction model associated with the preset risk variables, the method also includes: obtaining historical characteristic values ​​of the N target characteristics to obtain multiple groups of historical characteristic values; obtaining the risk coefficient of the preset risk variable corresponding to each group of historical characteristic values ​​to obtain multiple risk coefficients; determining each group of historical characteristic values ​​and the corresponding risk coefficient as a group of sample data to obtain multiple groups of sample data; training the neural network model through the multiple groups of sample data to obtain a risk prediction model.

[0123] Specifically, before performing the prediction operation, it is necessary to train the risk prediction model to ensure the accuracy of the risk prediction model. First, extract the characteristic values ​​of N target features at multiple historical moments from the blockchain database to form a data set containing all the historical data of the target features, and calculate the quantitative risk index corresponding to the preset risk variable corresponding to each set of historical characteristic values, that is, the risk coefficient, so as to determine the numerical correspondence between the historical financial characteristics and the risk coefficient, which is convenient for model understanding and learning.

[0124] Furthermore, each set of historical characteristic values ​​is combined with its corresponding risk coefficient to form a sample data set for training the neural network model, and the neural network model is iteratively trained using the constructed sample data set, and the model parameters are continuously adjusted until the model achieves satisfactory prediction performance on the training data. This enables the model to accurately predict the preset risk variables based on the characteristic values ​​of N target features, thereby improving the accuracy and efficiency of risk prediction.

[0125] This embodiment determines historical data and uses the data to train a neural network model, and ultimately obtains a risk prediction model associated with a preset risk variable. Through the collection of historical data, the calculation of risk coefficients, the construction of sample data, and the training of the model, it is ensured that the model can understand and learn the complex relationship between financial characteristics and risks, providing financial institutions with accurate risk assessment and prediction capabilities, and ensuring the accuracy and reliability of the risk prediction model.

[0126] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0127] Example 2

[0128] The embodiment of the present application also provides a risk prediction device. It should be noted that the risk prediction device of the embodiment of the present application can be used to execute the risk prediction method provided in the above embodiment. The risk prediction device provided in the embodiment of the present application is introduced below.

[0129] According to an embodiment of the present application, a device for implementing the above risk prediction method is also provided. Figure 3 is a schematic diagram of a risk prediction device provided according to Example 2 of the present application, such as Figure 3 As shown, the device comprises:

[0130] The first acquisition unit 31 is used to determine a preset risk variable and obtain a financial feature set of a target user from a blockchain database, wherein the financial feature set includes M financial features and a feature value of each financial feature, and M is a positive integer.

[0131] The selection unit 32 is used to select financial features associated with the preset risk variable from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set, to obtain N target features, where N is a positive integer.

[0132] The prediction unit 33 is used to input the characteristic values ​​of the N target characteristics into the risk prediction model associated with the preset risk variable to obtain the financial risk of the target user under the preset risk variable dimension.

[0133] In this embodiment, the preset risk variable is determined by the first acquisition unit 31, and the financial feature set of the target user is obtained from the blockchain database, wherein the financial feature set includes M financial features and the feature value of each financial feature, and M is a positive integer. The selection unit 32 selects the financial features associated with the preset risk variable from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set, and obtains N target features, wherein N is a positive integer. The prediction unit 33 inputs the feature values ​​of the N target features into the risk prediction model associated with the preset risk variable, and obtains the financial risk of the target user under the preset risk variable dimension. By obtaining the financial feature set of the user, and selecting the financial features associated with the preset risk variable to be predicted from the financial feature set, the target features are obtained, and then the target features are used to predict the financial risk corresponding to the preset risk variable, so as to achieve the purpose of selecting features according to different financial risks and accurately predicting the risk according to the features, thereby achieving the technical effect of improving the accuracy of financial risk prediction, and thus solving the technical problem of low accuracy of determining financial risks by artificial experience in related technologies.

[0134] Optionally, in the risk prediction device provided in the embodiment of the present application, the selection unit 32 includes: a first acquisition subunit, used to obtain the correlation and gain value between each financial feature in the financial feature set and the preset risk variable, and obtain M groups of parameter sets, wherein each group of parameter sets corresponds to the correlation and gain value of a financial feature; a second acquisition subunit, used to obtain the weights of the correlation and gain values, and perform weighted summation of the parameters in each group of parameter sets according to the weights to obtain M scores; a third acquisition subunit, used to obtain the number of parameters required for using the risk prediction model, and select financial features from the M financial features in descending order of scores according to the number of parameters to obtain N target features.

[0135] Optionally, in the risk prediction device provided in the embodiment of the present application, the first acquisition subunit includes: a first acquisition module, used to obtain, for any financial feature, historical feature values ​​of the financial feature and historical variable values ​​of a preset risk variable at multiple historical moments to obtain multiple sets of target numerical values; a first calculation module, used to calculate a first average value of the financial feature and a second average value of the preset risk variable based on multiple sets of target numerical values; a second calculation module, used to calculate the covariance between the financial feature and the preset risk variable based on multiple sets of target numerical values, the first average value and the second average value; a third calculation module, used to calculate the first standard deviation of the financial feature and the second standard deviation of the preset risk variable based on multiple sets of target numerical values, the first average value and the second average value; a first determination module, used to determine the correlation based on the first standard deviation, the second standard deviation and the covariance.

[0136] Optionally, in the risk prediction device provided in the embodiment of the present application, the first acquisition subunit includes: a second acquisition module, used to obtain, for any financial feature, historical feature values ​​of the financial feature and historical variable values ​​of a preset risk variable at multiple historical moments to obtain multiple sets of target numerical values; a fourth calculation module, used to calculate the entropy of the preset risk variable and the conditional entropy of the preset risk variable under the financial feature as a condition based on the multiple sets of target numerical values; a second determination module, used to determine the gain value based on the entropy and conditional entropy of the preset risk variable.

[0137] Optionally, in the risk prediction device provided in the embodiment of the present application, the first acquisition module includes: an input submodule, used to input the account information and private key of the target user into the blockchain database; a receiving submodule, used to receive multiple groups of initial data sets of the target user sent by the blockchain database when the private key is correct; a judgment submodule, used to judge whether there are candidate data sets with missing data in the multiple groups of initial data sets: a first calculation submodule, used to calculate the missing values ​​according to the financial characteristic values ​​contained in the candidate data sets when there are candidate data sets, and fill the missing values ​​into the multiple groups of initial data sets to obtain multiple groups of target numerical value sets.

[0138] Optionally, in the risk prediction device provided in the embodiment of the present application, the first calculation submodule includes: an acquisition submodule, used to obtain the position number of the missing value in the candidate data set, and obtain multiple first eigenvalues ​​with the smallest difference from the position number; a second calculation submodule, used to calculate the adjustment coefficient of the missing value based on the financial eigenvalue contained in the candidate data set; and a third calculation submodule, used to calculate the missing value based on the adjustment coefficient and multiple first eigenvalues.

[0139] Optionally, in the risk prediction device provided in the embodiment of the present application, before the characteristic values ​​of N target characteristics are input into the risk prediction model associated with the preset risk variables, the device also includes: a second acquisition unit, used to obtain historical characteristic values ​​of the N target characteristics to obtain multiple groups of historical characteristic values; a third acquisition unit, used to obtain the risk coefficient of the preset risk variable corresponding to each group of historical characteristic values ​​to obtain multiple risk coefficients; a determination unit, used to determine each group of historical characteristic values ​​and the corresponding risk coefficient as a group of sample data to obtain multiple groups of sample data; a training unit, used to train the neural network model through multiple groups of sample data to obtain a risk prediction model.

[0140] It should be noted here that the above-mentioned first acquisition unit 31, selection unit 32, and prediction unit 33 correspond to steps S201 to S203 in Example 1, and the two modules and the corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above-mentioned Example 1. It should be noted that the above-mentioned modules or units can be hardware components or software components stored in a memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n), and the above-mentioned modules can also be run in the computer terminal 10 provided in Example 1 as part of the device.

[0141] Example 3

[0142] An embodiment of the present application may provide an electronic device, Figure 4 is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 4 As shown, the electronic device may include: one or more ( Figure 4 (only one is shown) processor 1002, memory 1004, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0143] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0144] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: determine the preset risk variable, and obtain the financial feature set of the target user from the blockchain database, wherein the financial feature set includes M financial features and the feature value of each financial feature, and M is a positive integer; select the financial features associated with the preset risk variable from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set, and obtain N target features, wherein N is a positive integer; input the feature values ​​of the N target features into the risk prediction model associated with the preset risk variable, and obtain the financial risk of the target user under the preset risk variable dimension.

[0145] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: obtain the correlation and gain value between each financial feature in the financial feature set and the preset risk variable to obtain M groups of parameter sets, wherein each group of parameter sets corresponds to the correlation and gain value of a financial feature; obtain the weights of the correlation and gain values, and perform weighted summation of the parameters in each group of parameter sets according to the weights to obtain M scores; obtain the number of parameters required for using the risk prediction model, and select financial features from the M financial features in descending order of scores according to the number of parameters to obtain N target features.

[0146] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: for any financial feature, obtain the historical feature values ​​of the financial feature and the historical variable values ​​of the preset risk variable at multiple historical moments to obtain multiple sets of target numerical values; calculate the first average value of the financial feature and the second average value of the preset risk variable based on the multiple sets of target numerical values; calculate the covariance between the financial feature and the preset risk variable based on the multiple sets of target numerical values, the first average value and the second average value; calculate the first standard deviation of the financial feature and the second standard deviation of the preset risk variable based on the multiple sets of target numerical values, the first average value and the second average value; determine the correlation based on the first standard deviation, the second standard deviation and the covariance.

[0147] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: for any financial feature, obtain the historical feature values ​​of the financial feature and the historical variable values ​​of the preset risk variable at multiple historical moments to obtain multiple sets of target numerical values; calculate the entropy of the preset risk variable and the conditional entropy of the preset risk variable under the financial feature as the condition based on the multiple sets of target numerical values; determine the gain value based on the entropy of the preset risk variable and the conditional entropy.

[0148] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: input the account information and private key of the target user into the blockchain database; if the private key is correct, receive multiple sets of initial data sets of the target user sent by the blockchain database; determine whether there are candidate data sets with missing data in the multiple sets of initial data sets; if there are candidate data sets, calculate the missing values ​​based on the financial feature values ​​contained in the candidate data sets, and fill the missing values ​​into the multiple sets of initial data sets to obtain multiple sets of target numerical values.

[0149] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: obtain the position number of the missing value in the candidate data set, and obtain multiple first eigenvalues ​​with the smallest difference from the position number; calculate the adjustment coefficient of the missing value based on the financial eigenvalues ​​contained in the candidate data set; calculate the missing value based on the adjustment coefficient and multiple first eigenvalues.

[0150] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: obtain the historical feature values ​​of N target features to obtain multiple groups of historical feature values; obtain the risk coefficient of the preset risk variable corresponding to each group of historical feature values ​​to obtain multiple risk coefficients; determine each group of historical feature values ​​and the corresponding risk coefficient as a group of sample data to obtain multiple groups of sample data; train the neural network model through multiple groups of sample data to obtain a risk prediction model.

[0151] It can be understood by those skilled in the art that Figure 4 The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Figure 4 The structure of the electronic device is not limited. Figure 4 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 4 Different configurations are shown.

[0152] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0153] Example 4

[0154] The embodiment of the present application further provides a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the risk prediction method provided in the first embodiment.

[0155] Optionally, in this embodiment, the above storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0156] An embodiment of the present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing the steps of the risk prediction method.

[0157] An embodiment of the present application also provides a computer-readable storage medium, which includes a stored executable program, wherein when the executable program runs, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned risk prediction method.

[0158] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0159] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0160] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may 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 mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0161] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0162] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0163] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0164] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A risk prediction method, characterized in that: include: Determine a preset risk variable, and obtain a financial feature set of the target user from the blockchain database, wherein the financial feature set includes M financial features and a feature value of each financial feature, and M is a positive integer; Selecting financial features associated with the preset risk variable from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set to obtain N target features, where N is a positive integer; The characteristic values ​​of the N target characteristics are input into the risk prediction model associated with the preset risk variable to obtain the financial risk of the target user under the dimension of the preset risk variable.

2. The method according to claim 1, characterized in that According to the correlation and gain value between the preset risk variable and the financial feature set, a financial feature associated with the preset risk variable is selected from the financial feature set to obtain N target features including: Obtaining the correlation and gain value between each financial feature in the financial feature set and the preset risk variable to obtain M sets of parameter sets, wherein each set of parameter sets corresponds to the correlation and gain value of a financial feature; Obtaining weights of the correlation and the gain value, and performing weighted summation on the parameters in each parameter set according to the weights to obtain M scores; The number of parameters required for using the risk prediction model is obtained, and financial features are selected from the M financial features in descending order of the scores according to the number of parameters to obtain the N target features.

3. The method according to claim 1, characterized in that Obtaining the correlation between each financial feature in the financial feature set and the preset risk variable includes: For any financial feature, historical feature values ​​of the financial feature and historical variable values ​​of the preset risk variable at multiple historical moments are obtained to obtain multiple sets of target value sets; Calculating a first average value of the financial feature and a second average value of the preset risk variable according to the multiple sets of target value sets; Calculating the covariance between the financial feature and the preset risk variable according to the multiple sets of target value sets, the first average value, and the second average value; Calculate a first standard deviation of the financial feature and a second standard deviation of the preset risk variable according to the multiple sets of target value sets, the first average value, and the second average value; The correlation is determined based on the first standard deviation, the second standard deviation, and the covariance.

4. The method according to claim 1, characterized in that: Obtaining a gain value between each financial feature in the financial feature set and the preset risk variable includes: For any financial feature, historical feature values ​​of the financial feature and historical variable values ​​of the preset risk variable at multiple historical moments are obtained to obtain multiple sets of target value sets; Calculate the entropy of the preset risk variable and the conditional entropy of the preset risk variable under the financial feature as a condition according to the multiple sets of target numerical values; The gain value is determined according to the entropy of the preset risk variable and the conditional entropy.

5. The method according to claim 3 or 4, characterized in that: Obtaining historical characteristic values ​​of the financial characteristic and historical variable values ​​of the preset risk variable at multiple historical moments, and obtaining multiple sets of target value sets including: Inputting the target user's account information and private key into the blockchain database; When the private key is correct, receiving multiple sets of initial data of the target user sent by the blockchain database; Determine whether there is a candidate data set with missing data in the multiple groups of initial data sets; In the case where the candidate data set exists, the missing values ​​are calculated according to the financial feature values ​​included in the candidate data set, and the missing values ​​are filled into the multiple groups of initial data sets to obtain the multiple groups of target value sets.

6. The method according to claim 5, characterized in that Calculating missing values ​​according to the financial feature values ​​included in the candidate data set includes: Obtaining a position number of the missing value in the candidate data set, and obtaining a plurality of first eigenvalues ​​having the smallest difference with the position number; Calculating the adjustment coefficient of the missing value according to the financial feature value included in the candidate data set; The missing value is calculated according to the adjustment coefficient and the plurality of first eigenvalues.

7. The method according to claim 1, characterized in that Before inputting the characteristic values ​​of the N target characteristics into the risk prediction model associated with the preset risk variable, the method further includes: Obtaining historical feature values ​​of the N target features to obtain multiple groups of historical feature values; Obtain the risk coefficient of the preset risk variable corresponding to each group of historical characteristic values ​​to obtain multiple risk coefficients; Determine each set of historical characteristic values ​​and the corresponding risk coefficient as a set of sample data, and obtain multiple sets of sample data; The neural network model is trained using the multiple groups of sample data to obtain the risk prediction model.

8. A risk prediction device, characterized in that: include: A first acquisition unit is used to determine a preset risk variable and acquire a financial feature set of a target user from a blockchain database, wherein the financial feature set includes M financial features and a feature value of each financial feature, and M is a positive integer; a selection unit, configured to select financial features associated with the preset risk variable from the financial feature set according to the correlation and gain value between the preset risk variable and the financial feature set, to obtain N target features, where N is a positive integer; The prediction unit is used to input the characteristic values ​​of the N target characteristics into the risk prediction model associated with the preset risk variable to obtain the financial risk of the target user under the dimension of the preset risk variable.

9. A computer program product comprising computer instructions, characterized in that When the computer instructions are executed by a processor, the steps of the risk prediction method described in any one of claims 1 to 7 are implemented.

10. An electronic device, characterized in that: include: A memory storing an executable program; A processor is used to run the program, wherein the risk prediction method described in any one of claims 1 to 7 is executed when the program is run.

Images