Risk control strategy automatic generation method and device, equipment and medium

Abstract

The invention relates to the technical field of intelligent risk control, and discloses a risk control strategy automatic generation method, device, equipment and medium, and the method comprises the steps: selecting a sample and cleaning sample data, enabling the cleaned sample to take a user ID as a main key, and enabling a field in the cleaned sample data to comprise variable features; training a decision-making tree model by using the sample data based on a random forest algorithm, adjusting parameters of the decision-making tree model, training nodes of the decision-making tree model, and determining optimal variable characteristics to perform left and right sub-tree division of the decision-making tree; a tree and a corresponding leaf node are generated based on the decision tree model, variable features in the leaf node are converted into new variable features based on business logic and IV value selection, and the new variable features serve as an updating iteration strategy of the decision tree model. The efficiency of automatically generating the strategies based on the random forest model is high, the strategies are screened more efficiently based on the service logic, the IV value and the stability of the data, and the random forest generally can provide higher prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent risk control technology and natural language processing technology, and in particular to a method, device, equipment and medium for automatically generating risk control strategies. Background Art

[0002] Faced with massive amounts of data, how to quickly and efficiently mine effective risk control strategies is a very difficult topic. In business security, strategy risk control is a very important link, but traditional methods mainly rely on manual experience to discover high-risk strategies, and manual methods have two limitations: 1. Manual mining efficiency is low, and the person mining is required to have rich business experience. 2. The adequacy of the mining strategy is relatively limited. Faced with massive amounts of data, the amount that can be processed manually is limited, and the accuracy is low, and the mining strategy is not complete and comprehensive. Summary of the invention

[0003] The present invention provides a method, device, computer equipment and medium for automatically generating risk control strategies to solve the technical problems of high threshold, low efficiency and low accuracy in traditional manual processing of mining risks in massive data.

[0004] In a first aspect, a method for automatically generating a risk control strategy is provided, comprising:

[0005] Select samples and clean the sample data. The cleaned samples use user ID as the primary key. The fields in the cleaned sample data contain variable features. The number of candidate questions is used as matching questions.

[0006] A matching answer corresponding to the matching question is selected to respond to the user question.

[0007] In a second aspect, a device for automatically generating a risk control strategy is provided, comprising:

[0008] A processing module is used to select samples and clean sample data. The cleaned samples use user ID as the primary key, and the fields in the cleaned sample data contain variable features.

[0009] The training module is used to train the decision tree model with sample data based on the random forest algorithm, adjust the parameters of the decision tree model, train the nodes of the decision tree model, and confirm the optimal variable features for left and right subtree division of the decision tree;

[0010] The iteration module is used to generate a tree and corresponding leaf nodes based on the decision tree model, convert the variable features in the leaf nodes into new variable features based on business logic and IV value selection, and use the new variable features as the update iteration strategy of the decision tree model.

[0011] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for automatically generating risk control strategies when executing the computer program.

[0012] In a fourth aspect, a computer-readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for automatically generating risk control strategies are implemented.

[0013] In the scheme implemented by the above-mentioned risk control strategy automatic generation method, device, computer equipment and storage medium, samples can be selected and sample data can be cleaned. The cleaned sample uses the user ID as the primary key, and the fields in the cleaned sample data contain variable features; the sample data is trained with a decision tree model based on the random forest algorithm, the parameters of the decision tree model are adjusted, and the nodes of the decision tree model are trained to confirm the optimal variable features for left and right subtree division of the decision tree; a tree and corresponding leaf nodes are generated based on the decision tree model, and the variable features in the leaf nodes are converted into new variable features based on business logic and IV value selection, and the new variable features are used as the update iteration strategy of the decision tree model. Automatically generating strategies based on the random forest model is highly efficient, and hundreds or thousands of alternative strategies can be generated each time. It is more efficient to screen strategies based on business logic, IV values, and data stability. Random forests can usually provide higher prediction accuracy. The random forest model can sample horizontally and vertically, and each tree generated by the random forest is independent. The random forest has a better tolerance for noise and outliers in the data because it relies on the prediction results of multiple trees rather than a single tree. The random forest can evaluate the importance of each feature to the prediction results, which helps with feature selection and data understanding. By integrating multiple trees, random forests reduce the risk of overfitting and generate more comprehensive strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0015] Figure 1 This is a schematic diagram of an application environment of a method for automatically generating risk control strategies in an embodiment of the present invention;

[0016] Figure 2 This is a flow chart of a method for automatically generating a risk control strategy in one embodiment of the present invention;

[0017] Figure 3It is a structural schematic diagram of a device for automatically generating risk control strategies in one embodiment of the present invention;

[0018] Figure 4 is a schematic diagram of a structure of a computer device in one embodiment of the present invention;

[0019] Figure 5 It is another structural schematic diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] The method for automatically generating risk control strategies provided by the embodiment of the present invention can be applied in the following aspects: Figure 1 In the application environment, the client communicates with the server through the network. The server can receive user questions through the client, and perform question recall processing from the preset standard question library according to the user questions, so as to select samples and clean the sample data. The cleaned sample uses the user ID as the primary key, and the fields in the cleaned sample data contain variable features; the sample data is trained with the decision tree model based on the random forest algorithm, the parameters of the decision tree model are adjusted, and the nodes of the decision tree model are trained to confirm the optimal variable features for the left and right subtree division of the decision tree; the tree and the corresponding leaf nodes are generated based on the decision tree model, and the variable features in the leaf nodes are converted into new variable features based on the business logic and IV value selection, and the new variable features are used as the update iteration strategy of the decision tree model. Automatically generating strategies based on the random forest model is highly efficient, and hundreds or thousands of alternative strategies can be generated each time. It is more efficient to screen strategies based on business logic, IV value, and data stability. Random forests can usually provide higher prediction accuracy; random forest models can sample horizontally and vertically, and each tree generated by random forests is independent. Random forests have better tolerance for noise and outliers in data because they rely on the prediction results of multiple trees, not a single tree; random forests can evaluate the importance of each feature to the prediction results, which helps feature selection and data understanding; by integrating multiple trees, random forests reduce the risk of overfitting, and the generated strategies are more comprehensive. The server can be implemented with an independent server or a server cluster consisting of multiple servers. The present invention is described in detail below through specific embodiments.

[0022] See also Figure 2 As shown, Figure 2A flow chart of a method for automatically generating a risk control strategy provided by an embodiment of the present invention includes the following steps:

[0023] S10: Select samples and clean sample data. The cleaned samples use user ID as the primary key. The fields in the cleaned sample data contain variable features.

[0024] Before training, data preprocessing is also required, which is the process of processing and organizing the raw data to ensure data quality, handle outliers, etc., reduce the error of the final result, and improve the quality and availability of the data.

[0025] Data cleaning steps: remove duplicate data, identify and delete duplicate records; handle missing values, fill in missing values ​​or delete records containing missing values; unify data formats, unify data formats, such as date, time or numerical units; correct data errors, correct errors and inconsistencies in data, such as spelling errors or incorrect classifications; handle outliers, detect and handle outliers or outliers in data; data consistency check, ensure the consistency of data between different data sources or tables; data conversion, convert data into a format suitable for analysis, such as normalization, encoding categorical variables, etc.

[0026] S20: Based on the random forest algorithm, the sample data is trained into a decision tree model, the parameters of the decision tree model are adjusted, and the nodes of the decision tree model are trained to confirm the optimal variable features for dividing the decision tree into left and right subtrees;

[0027] Specifically, multiple different training sample data sets are extracted from the sample data by self-sampling; variable features are sampled a preset number of times with replacement sampling in the training sample data set to form a new sub-training set; a preset number of variable features are randomly selected, where the preset number is less than the number of variable features of the sample data; and a new sub-training set and a preset number of variable features are used for training and learning to obtain a complete decision tree model.

[0028] Train a decision tree model based on sample data, adjust the parameters of the decision tree model, and optimize the stability and accuracy of the decision tree model. Because the training of the decision tree model is used to generate strategies, the tree depth and the number of leaf nodes of each decision tree are generally limited, and the decision tree model and parameters are saved. Self-service sampling: Extract multiple different training sample data sets from the original sample data through self-service sampling. Self-service sampling allows samples in the sample data set to recur, which means that the size of each training sample data set is the same as the original sample data set, but contains some repeated samples.

[0029] Random feature selection: At each split node of each decision tree, a portion of variable features are randomly selected instead of considering all possible variable features. This further increases the diversity of the decision tree model and helps reduce overfitting.

[0030] The introduction of the decision tree model is as follows: Input is a sample data set

[0031] D={(x1,y1),(x2,y2),...(x m ,y m )}, the number of iterations of the weak classifier is T; the output is the final strong classifier f(x).

[0032] It should be noted that after the step of sampling a preset number of variable features in the training sample data set with replacement to form a new sub-training set, the method further includes: when training the node of the decision tree model, randomly sampling a preset number of variable features from all variable features on the node; and selecting the optimal variable feature from the randomly sampled preset number of variable features to divide the decision tree into left and right subtrees. For t=1,2...,T:

[0033] a) Perform random sampling on the training set for the tth time, a total of m times, to obtain a sampling set D containing m samples t

[0034] b) Using sampling set D t Train the tth decision tree model G t (x), when training the nodes of the decision tree model, select a part of the sample features from all the sample features on the node, and select the best feature from these randomly selected sample features to divide the decision tree into left and right subtrees.

[0035] If it is a classification algorithm prediction, input the training sample data set and the preset number of variable features; based on the regression algorithm, the regression results obtained by regressing the variable features for the preset number of times are averaged to obtain the final new sub-training set. Then the category or one of the categories with the most votes from T weak learners is the final category. If it is a regression algorithm, the value obtained by averaging the regression results obtained by T weak learners is the final model output.

[0036] Prediction and aggregation: For classification problems, each decision tree gives a prediction, and the final prediction result is the majority vote of the prediction results of all decision trees; for regression problems, it is the average of the prediction results of all decision trees.

[0037] Specifically, random selection of data: construct a new sub-training set from the training sample data set by sampling with replacement, and the data of the new sub-training set is the same as the training sample data set; construct a sub-decision tree based on the new sub-training set, and store the data of the new sub-training set in each sub-decision tree. First, take a sampling with replacement from the original data set to construct a sub-dataset, and the data of the sub-dataset is the same as the original data set. Elements of different sub-datasets can be repeated, and elements in the same sub-dataset can also be repeated. Second, use the sub-dataset to construct a sub-decision tree, put this data into each sub-decision tree, and each sub-decision tree outputs a result. Finally, if there is new data that needs to be classified through random forest, the output result of random forest can be obtained by voting on the judgment results of the sub-decision tree. Assuming that there are 3 sub-decision trees in the random forest, the classification results of 2 sub-decision trees are class A, and the classification result of 1 sub-decision tree is class B, then the classification result of random forest is class A.

[0038] Random selection of candidate features: At each splitting node of the training sample data set, a preset number of variable features are randomly selected; the best variable features are selected from the preset number of variable features. Similar to the random selection of data sets, each splitting process of the subtree in the random forest does not use all the candidate features, but randomly selects certain features from all the candidate features, and then selects the best features from the randomly selected features. This allows the decision trees in the random forest to be different from each other, improves the diversity of the system, and thus improves the classification performance.

[0039] The advantages of the random forest algorithm include: high accuracy: due to the integration of multiple decision trees, random forests are generally able to provide higher prediction accuracy; robustness: random forests have better tolerance to noise and outliers in the data because it relies on the prediction results of multiple trees rather than a single tree; feature importance assessment: random forests can evaluate the importance of each feature to the prediction results, which helps with feature selection and data understanding; anti-overfitting ability: by integrating multiple trees, random forests reduce the risk of overfitting.

[0040] S30: A tree and corresponding leaf nodes are generated based on the decision tree model. The variable features in the leaf nodes are converted into new variable features based on business logic and IV value selection, and the new variable features are used as an update iteration strategy for the decision tree model.

[0041] Specifically, some trees and corresponding leaf nodes are generated based on the random forest model, and the data is converted to convert the leaf nodes into variables of 0 and 1. The variable features with the best information value are selected from the leaf nodes based on business logic and IV value, and the variable features with the best information value are used as new variable features, and the new variable features are used as the update iteration strategy of the decision tree model.

[0042] IV stands for Information Value, which is generally used to screen input features and to measure the predictive ability of features for the model. The IV value is the symmetric KL distance, which can measure the difference in category distribution, the bad sample distribution and the good sample distribution from the category perspective, reflecting the predictive ability of the input variable for the label. The larger the value, the better.

[0043]

[0044] In this embodiment, petal_width>0.75 has strong predictive power, petal_width<=1.65 has medium predictive power, and petal_length<=4.85 has weak predictive power.

[0045] The business logic can be transformed into a new variable feature tree4 (0, 1 variable). If the business logic is satisfied, the value is 1, otherwise it is 0. In this way, all leaf nodes are transformed into new variable features. The business logic of these leaf nodes is the alternative strategy. The strategy is selected based on the business logic and IV value, so that the strategy is automatically generated through the model.

[0046] It can be seen that in the above scheme, for the iteration of intelligent risk control data, first, samples are selected and the sample data is cleaned. The cleaned samples use user ID as the primary key, and the fields in the cleaned sample data contain variable features; the sample data is used to train the decision tree model based on the random forest algorithm, the parameters of the decision tree model are adjusted, and the nodes of the decision tree model are trained to confirm the optimal variable features for the left and right subtree division of the decision tree; the tree and the corresponding leaf nodes are generated based on the decision tree model, and the variable features in the leaf nodes are converted into new variable features based on business logic and IV value selection, and the new variable features are used as the update iteration strategy of the decision tree model. Automatically generating strategies based on the random forest model is highly efficient, and hundreds or thousands of alternative strategies can be generated each time. It is more efficient to screen strategies based on business logic, IV values, and data stability. Random forests can usually provide higher prediction accuracy. The random forest model can sample horizontally and vertically, and each tree generated by the random forest is independent. The random forest has a better tolerance for noise and outliers in the data because it relies on the prediction results of multiple trees rather than a single tree. The random forest can evaluate the importance of each feature to the prediction results, which helps with feature selection and data understanding. By integrating multiple trees, random forests reduce the risk of overfitting and generate more comprehensive strategies.

[0047] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0048] In one embodiment, a device for automatically generating a risk control strategy is provided, and the device for automatically generating a risk control strategy corresponds one-to-one to the method for automatically generating a risk control strategy in the above embodiment. Figure 3 As shown, the risk control strategy automatic generation device includes a processing module 101, a training module 102, and an iteration module 103. Each functional module is described in detail as follows:

[0049] Processing module 101, used to select samples and clean sample data, the cleaned samples use user ID as primary key, and the fields in the cleaned sample data contain variable features;

[0050] The training module 102 is used to train the decision tree model with the sample data based on the random forest algorithm, adjust the parameters of the decision tree model, and train the nodes of the decision tree model to determine the optimal variable features for dividing the decision tree into left and right subtrees;

[0051] The iteration module 103 is used to generate a tree and corresponding leaf nodes based on the decision tree model, convert the variable features in the leaf nodes into new variable features based on business logic and IV value selection, and use the new variable features as an update iteration strategy for the decision tree model.

[0052] In one embodiment, the training module 102 is specifically configured to:

[0053] Extract multiple different training sample data sets from the sample data by self-sampling;

[0054] Based on the variable features in the training sample data set, a new sub-training set is formed by sampling a preset number of times with replacement sampling;

[0055] Randomly select a preset number of variable features, where the preset number is less than the number of variable features in the sample data;

[0056] Using a new sub-training set and a preset number of variable features, training and learning are performed to obtain a complete decision tree model.

[0057] In one embodiment, the training module 102 is further configured to:

[0058] Input the training sample data set and the preset number of variable features;

[0059] Based on the regression algorithm, the regression results obtained by regressing the variable features for a preset number of times are averaged to obtain the final new sub-training set.

[0060] In one embodiment, the training module 102 is further configured to:

[0061] Construct a new sub-training set from the training sample data set by sampling with replacement, and the data of the new sub-training set is the same as that of the training sample data set;

[0062] A sub-decision tree is constructed based on the new sub-training set, and the data of the new sub-training set is stored in each sub-decision tree.

[0063] In one embodiment, the training module 102 is further configured to:

[0064] When training a node of a decision tree model, randomly sample a preset number of variable features from all variable features on the node;

[0065] The optimal variable feature is selected from the variable features randomly sampled a preset number of times to divide the decision tree into left and right subtrees.

[0066] In one embodiment, the training module 102 is further configured to:

[0067] At each split node of the training sample data set, a preset number of variable features are randomly selected;

[0068] Select the best variable feature from a preset number of variable features.

[0069] In one embodiment, the iteration module 103 is specifically configured to:

[0070] Based on business logic and IV value selection, the variable features with the best information value in the leaf nodes are selected, the variable features with the best information value are used as new variable features, and the new variable features are used as the update iteration strategy of the decision tree model.

[0071] The present invention provides an automatic generation device for risk control strategies. First, samples are selected and sample data is cleaned. The cleaned samples use user IDs as primary keys, and the fields in the cleaned sample data contain variable features. A decision tree model is trained with the sample data based on a random forest algorithm, the parameters of the decision tree model are adjusted, and the nodes of the decision tree model are trained to confirm the optimal variable features for left and right subtree division of the decision tree. A tree and corresponding leaf nodes are generated based on the decision tree model, and the variable features in the leaf nodes are converted into new variable features based on business logic and IV value selection, and the new variable features are used as update and iteration strategies for the decision tree model. Automatically generating strategies based on the random forest model is highly efficient, and hundreds or even thousands of alternative strategies can be generated each time. It is more efficient to screen strategies based on business logic, IV values, and data stability. Random forests can usually provide higher prediction accuracy. The random forest model can sample horizontally and vertically, and each tree generated by the random forest is independent. The random forest has a better tolerance for noise and outliers in the data because it relies on the prediction results of multiple trees rather than a single tree. The random forest can evaluate the importance of each feature to the prediction results, which helps with feature selection and data understanding. By integrating multiple trees, the random forest reduces the risk of overfitting, and the generated strategies are more comprehensive.

[0072] For the specific definition of the risk control strategy automatic generation device, please refer to the definition of the risk control strategy automatic generation method above, which will not be repeated here. Each module in the above-mentioned risk control strategy automatic generation device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0073] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or 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 the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the service side of a method for automatically generating risk control strategies.

[0074] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. 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 and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it realizes the functions or steps of the client side of a method for automatically generating risk control strategies

[0075] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0076] Select samples and clean the sample data. The cleaned samples use user ID as the primary key. The fields in the cleaned sample data contain variable features. The number of candidate questions is used as matching questions.

[0077] A matching answer corresponding to the matching question is selected to respond to the user question.

[0078] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0079] Select samples and clean the sample data. The cleaned samples use user ID as the primary key. The fields in the cleaned sample data contain variable features. The number of candidate questions is used as matching questions.

[0080] A matching answer corresponding to the matching question is selected to respond to the user question.

[0081] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0082] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed 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, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0083] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0084] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for automatically generating a risk control strategy, characterized in that: include: Select samples and clean the sample data. The cleaned samples use user ID as the primary key. The fields in the cleaned sample data contain variable features. Based on the random forest algorithm, the sample data is used to train the decision tree model, the parameters of the decision tree model are adjusted, and the nodes of the decision tree model are trained to confirm the optimal variable features for the left and right subtree division of the decision tree; A tree and corresponding leaf nodes are generated based on the decision tree model. The variable features in the leaf nodes are converted into new variable features based on business logic and IV value selection, and the new variable features are used as the update iteration strategy of the decision tree model.

2. The method for automatically generating risk control strategies according to claim 1, characterized in that: The steps of training a decision tree model with sample data based on a random forest algorithm, adjusting the parameters of the decision tree model, training the nodes of the decision tree model, and confirming the optimal variable features for dividing the decision tree into left and right subtrees include: Extract multiple different training sample data sets from the sample data by self-sampling; Based on the variable features in the training sample data set, a new sub-training set is formed by sampling a preset number of times with replacement sampling; Randomly select a preset number of variable features, where the preset number is less than the number of variable features in the sample data; Using a new sub-training set and a preset number of variable features, training and learning are performed to obtain a complete decision tree model.

3. The method for automatically generating risk control strategies according to claim 2, characterized in that: The step of sampling a variable feature with replacement in the training sample data set for a preset number of times to form a new sub-training set comprises: Input the training sample data set and the preset number of variable features; Based on the regression algorithm, the regression results obtained by regressing the variable features for a preset number of times are averaged to obtain the final new sub-training set.

4. The method for automatically generating risk control strategies according to claim 3, characterized in that: The step of forming a new sub-training set by sampling the variable features of the training sample data set a preset number of times with replacement sampling also includes: Construct a new sub-training set from the training sample data set by sampling with replacement, and the data of the new sub-training set is the same as that of the training sample data set; A sub-decision tree is constructed based on the new sub-training set, and the data of the new sub-training set is stored in each sub-decision tree.

5. The method for automatically generating risk control strategies according to claim 2, characterized in that: After the step of sampling the variable features of the training sample data set with replacement for a preset number of times to form a new sub-training set, the method further comprises: When training a node of a decision tree model, randomly sample a preset number of variable features from all variable features on the node; The optimal variable feature is selected from the variable features randomly sampled a preset number of times to divide the decision tree into left and right subtrees.

6. The method for automatically generating risk control strategies according to claim 2, characterized in that: The step of randomly selecting a preset number of variable features comprises: At each split node of the training sample data set, a preset number of variable features are randomly selected; Select the best variable feature from a preset number of variable features.

7. The method for automatically generating risk control strategies according to claim 1, characterized in that: The steps of generating a tree and corresponding leaf nodes based on the decision tree model, converting variable features in the leaf nodes into new variable features based on business logic and IV value selection, and using the new variable features as an update iteration strategy for the decision tree model include: Based on business logic and IV value selection, the variable features with the best information value in the leaf nodes are selected, the variable features with the best information value are used as new variable features, and the new variable features are used as the update iteration strategy of the decision tree model.

8. A device for automatically generating risk control strategies, characterized in that: include: A processing module is used to select samples and clean sample data. The cleaned samples use user ID as the primary key, and the fields in the cleaned sample data contain variable features. The training module is used to train the decision tree model with sample data based on the random forest algorithm, adjust the parameters of the decision tree model, train the nodes of the decision tree model, and confirm the optimal variable features for left and right subtree division of the decision tree; The iteration module is used to generate a tree and corresponding leaf nodes based on the decision tree model, convert the variable features in the leaf nodes into new variable features based on business logic and IV value selection, and use the new variable features as the update iteration strategy of the decision tree model.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for automatically generating risk control strategies as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for automatically generating risk control strategies as described in any one of claims 1 to 7 are implemented.

Images