Credit risk level assessment method, device and equipment and storage medium

Abstract

The invention discloses a credit risk level assessment method, device and equipment and a storage medium, and belongs to the field of artificial intelligence, and the method comprises the steps: obtaining to-be-assessed credit data of a target object; inputting the to-be-evaluated credit data into the target evaluation model to obtain a credit risk level corresponding to the target object; wherein the target evaluation model is obtained by training a credit risk level evaluation model based on an Adaboost algorithm based on a sample credit data set. According to the invention, the accuracy of user credit risk assessment is improved, and the credit risk in the credit business of banks and other financial institutions is reduced. Meanwhile, credit approvers do not need to check the loan application materials of the borrowers one by one, so that the loan approval efficiency of financial institutions such as banks is improved, and the business handling time of the borrowers is shortened.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a credit risk rating assessment method, device, equipment and storage medium. Background Art

[0002] In the financial industry, when a user applies for a loan from a financial institution, banks and other financial institutions usually assess the user's credit risk based on credit data such as the user's credit information and economic status. For banks and other financial institutions, accurately assessing the user's credit risk is crucial to maintaining stability and profitability.

[0003] However, existing credit risk rating assessment models are often unable to accurately assess users' credit risks due to quality issues with model training data, thereby increasing the credit risk in the credit business of banks and other financial institutions. Summary of the invention

[0004] The present invention provides a credit risk level assessment method, device, equipment and storage medium to improve the accuracy of user credit risk assessment and reduce the credit risk in the credit business of financial institutions such as banks.

[0005] According to one aspect of the present invention, a credit risk rating assessment method is provided, the method comprising:

[0006] Obtain the credit data of the target object to be evaluated;

[0007] The credit data to be evaluated is input into the target evaluation model to obtain the credit risk level corresponding to the target object; wherein the target evaluation model is obtained by training a credit risk level evaluation model based on the Adaboost algorithm based on a sample credit data set.

[0008] According to another aspect of the present invention, there is provided a credit risk level assessment device, the device comprising:

[0009] A module for acquiring credit data to be evaluated, used to acquire credit data to be evaluated of a target object;

[0010] The credit risk rating assessment module is used to input the credit data to be assessed into the target assessment model to obtain the credit risk rating corresponding to the target object; wherein the target assessment model is obtained by training the credit risk rating assessment model based on the Adaboost algorithm based on the sample credit data set.

[0011] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0012] at least one processor; and

[0013] a memory communicatively connected to at least one processor; wherein,

[0014] The memory stores a computer program executable by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the credit risk rating assessment method of any embodiment of the present invention.

[0015] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the credit risk rating assessment method of any embodiment of the present invention when executed.

[0016] According to another aspect of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the credit risk rating assessment method of any embodiment of the present invention is implemented.

[0017] The technical solution of the embodiment of the present invention obtains the credit data to be evaluated of the target object; inputs the credit data to be evaluated into the target evaluation model to obtain the credit risk level corresponding to the target object; wherein the target evaluation model is obtained by training the credit risk level evaluation model based on the Adaboost algorithm based on the sample credit data set. The above technical solution processes the credit data to be evaluated of the target object through the target evaluation model, thereby obtaining the credit risk level corresponding to the target object, improving the accuracy of the credit risk assessment of the user and reducing the credit risk in the credit business of banks and other financial institutions. At the same time, there is no need for credit approval personnel to review the loan application materials of the borrower one by one, thereby improving the loan approval efficiency of banks and other financial institutions and shortening the business processing time of the borrower.

[0018] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 is a flow chart of a credit risk rating assessment method provided according to Embodiment 1 of the present invention;

[0021] Figure 2is a flow chart of a credit risk rating assessment method provided according to Embodiment 2 of the present invention;

[0022] Figure 3 is a flow chart of a credit risk level assessment method provided according to Embodiment 3 of the present invention;

[0023] Figure 4 This is a schematic diagram of the structure of a credit risk level assessment device provided according to a fourth embodiment of the present invention;

[0024] Figure 5 It is a schematic diagram of the structure of an electronic device for implementing the credit risk level assessment method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "target", "sample", "first" and "second" in the specification and claims of the present invention 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 interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] It should be noted that the information collected in the present invention, such as the target object's credit data to be evaluated and the sample credit data set, is 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 the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0028] Embodiment 1

[0029] Figure 1This is a flowchart of a credit risk rating assessment method provided in Embodiment 1 of the present invention. This embodiment is applicable to the situation of assessing the credit risk rating of customers of financial institutions. The method can be performed by a credit risk rating assessment device, which can be implemented in the form of hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes:

[0030] S101. Obtain credit data to be evaluated of the target object.

[0031] The target object refers to the client of the financial institution who borrows money from the financial institution, referred to as the borrower. The credit data to be evaluated refers to the credit data required to evaluate the credit risk level corresponding to the target object; optionally, the credit data to be evaluated includes but is not limited to the personal information, financial information, credit history and social activity information of the target object. The credit history refers to a series of credit records and historical data generated by the target object in past lending activities, including the repayment record, loan limit and credit card usage of the target object.

[0032] Specifically, the target object's credit data to be evaluated can be obtained based on a preset data acquisition method. Exemplarily, the target object's credit data to be evaluated can be obtained from the database of the financial institution and the third-party data platform associated with the financial institution, using the target object's object identifier as an index. The object identifier is used to uniquely identify the target object; optionally, the object identifier can be in the form of numbers, letters, or numbers plus letters. The third-party data platform includes but is not limited to credit reporting agencies, social network platforms, and financial data service providers.

[0033] S102: Input the credit data to be evaluated into the target evaluation model to obtain the credit risk level corresponding to the target object.

[0034] The target assessment model is obtained by training a credit risk rating assessment model based on the Adaboost algorithm based on a sample credit data set. The target assessment model refers to a trained credit risk rating assessment model, which is used to assess the credit risk rating corresponding to the target object. The credit risk rating is used to characterize the degree of credit risk of the target object. The sample credit data set refers to the data set required for training the credit risk rating assessment model; optionally, the sample credit data set consists of sample credit data of a large number of sample objects. The sample object refers to an object randomly selected when training the credit risk rating assessment model. The sample credit data refers to the credit data of the sample object in a historical time period; optionally, the sample credit data includes but is not limited to the personal information, financial information, credit history, credit risk rating and social activity information of the sample object.

[0035] Among them, the Adaboost (Adaptive Boosting) algorithm is an integrated learning method based on the Boosting strategy. Its core idea is to train different weak classifiers for the same training set, and then combine these weak classifiers to form a stronger final classifier (i.e., a strong classifier). Among them, a weak classifier refers to a classifier whose classification accuracy is only slightly better than random guessing. After each round of iteration, the Adaboost algorithm updates the weights of the samples in the training set and the weights of the weak classifiers, and finally fuses the classifiers obtained from each training to obtain the final decision classifier.

[0036] Specifically, the credit data to be evaluated is input into the target evaluation model, and after being processed by the target evaluation model, the credit risk level corresponding to the target object is obtained.

[0037] The technical solution of the embodiment of the present invention obtains the credit data to be evaluated of the target object; inputs the credit data to be evaluated into the target evaluation model to obtain the credit risk level corresponding to the target object; wherein the target evaluation model is obtained by training the credit risk level evaluation model based on the Adaboost algorithm based on the sample credit data set. The above technical solution processes the credit data to be evaluated of the target object through the target evaluation model, thereby obtaining the credit risk level corresponding to the target object, improving the accuracy of the credit risk assessment of the user and reducing the credit risk in the credit business of banks and other financial institutions. At the same time, there is no need for credit approval personnel to review the loan application materials of the borrower one by one, thereby improving the loan approval efficiency of banks and other financial institutions and shortening the business processing time of the borrower.

[0038] Embodiment 2

[0039] Figure 2 This is a flowchart of a credit risk rating assessment method provided in Example 2 of the present invention. Based on the above example, this example further optimizes the "determination of the target assessment model" and provides an optional implementation plan. It should be noted that for the parts not described in detail in the example of the present invention, reference can be made to the relevant descriptions of other examples. Figure 2 As shown, the method includes:

[0040] S201. Obtain a sample credit data set.

[0041] S202: Perform feature selection on the sample credit data set to obtain a key credit data set.

[0042] Among them, the key credit data set refers to a data set consisting of data in the sample credit data set that has a greater impact on the credit risk level assessment.

[0043] Specifically, the first data feature and the second data feature in the sample credit data set can be determined; based on the Pearson correlation coefficient, the correlation between each sub-data feature in the second data feature and the first data feature is calculated; according to the correlation between each sub-data feature in the second data feature and the first data feature, feature selection is performed on the sample credit data set to obtain the key credit data set.

[0044] Among them, the Pearson correlation coefficient is a statistic that measures the strength of the linear relationship between variables. Its value range is between -1 and 1, which can reflect the degree of correlation between two variables. If the correlation coefficient is close to 1, it means that there is a completely positive linear relationship between the two variables; if the correlation coefficient is close to -1, it means that there is a completely negative linear relationship between the two variables; if the correlation coefficient is close to 0, it means that there is no linear relationship between the two variables.

[0045] The first data feature refers to the credit risk level in the sample credit data set, and the second data feature refers to the data feature in the sample credit data set other than the first data feature.

[0046] More specifically, the first data feature and the second data feature in the sample credit data set can be determined according to actual business needs; then, for the i-th sub-data feature X in the second data feature, i , based on the Pearson correlation coefficient, the i-th sub-data feature X in the second data feature is calculated by the following formula: i Correlation with the first data feature:

[0047]

[0048] Wherein, M represents the total number of eigenvalues. It should be noted that the total number of eigenvalues ​​corresponding to each sub-data feature in the second data feature is the same, and is equal to the total number of eigenvalues ​​corresponding to the first data feature. j Represents X i The corresponding j-th eigenvalue is, Represents X i The average value of the corresponding M eigenvalues, y j represents the jth eigenvalue corresponding to the first data feature, represents the average value of the M eigenvalues ​​corresponding to the first data feature, r i Represents the i-th sub-data feature X in the second data feature i Similarly, the correlation between each sub-data feature in the second data feature and the first data feature can be obtained, which are recorded as r1, r2, ... and r n . Wherein, n represents the total number of sub-data features in the second data feature.

[0049] Afterwards, the correlation between each sub-data feature and the first data feature is compared with the correlation threshold, and the sub-data features with correlation greater than the correlation threshold and the first data feature are saved in an empty data set, thereby obtaining a key credit data set. The correlation threshold can be preset according to actual business needs, and the embodiment of the present invention does not specifically limit it.

[0050] It can be understood that by performing feature selection on the data features in the sample credit data set based on the Pearson correlation coefficient, only the data features in the sample credit data set that are highly correlated with the credit risk level are retained, thereby reducing the amount of data for model training data and improving the data quality of model training data, thereby reducing the consumption of computing resources and improving data processing efficiency during model training.

[0051] S203: construct a training sample set according to the key credit data set, and train a credit risk rating assessment model based on the Adaboost algorithm according to the training sample set to obtain a target assessment model.

[0052] The training sample set refers to the data set actually used to train the credit risk rating assessment model.

[0053] Specifically, the data in the key credit data set can be processed based on the data processing model to obtain a training sample set output by the data processing model; wherein the data processing model is used to solve the data missing problem and the data imbalance problem in the data set; then, the total number of samples of the training samples in the training sample set can be determined by a statistical algorithm; the total number of classifiers of the basic classifier in the Adaboost algorithm can be determined by a cross-validation strategy; wherein the cross-validation strategy can be a K-fold cross-validation strategy or a leave-one-out cross-validation strategy; the initial weight value of each training sample in the training sample set is determined according to the total number of samples, that is, the inverse of the total number of samples is used as the initial weight value of each training sample in the training sample set, so that the initial weight value distribution of the training sample set can be obtained:

[0054]

[0055] Where D1(i) represents the initial weight value distribution of the training sample set, and also represents the weight value distribution of the training sample set during the first training; i represents the i-th training sample in the training sample set, and i = 1, 2, …, m; m represents the total number of training samples in the training sample set; w m It indicates the weight value of the mth training sample in the training sample set. It should be noted that the basic classifiers in the Adaboost algorithm are all decision tree stumps.

[0056] After that, all the basic classifiers in the Adaboost algorithm are trained according to the training sample set that assigns a weight value to each training sample, and then the basic classifiers are combined to obtain the target evaluation model. More specifically, during the kth training, the training sample set T obtained after the k-1th training is k-1 Output to the kth basic classifier in the Adaboost algorithm, and get the kth basic classifier for the training sample set T k-1 The classification result of each training sample in ; According to the obtained classification results and the training sample set T k-1 The actual classification results corresponding to each training sample in the Adaboost algorithm determine the classification error rate of the kth basic classifier:

[0057]

[0058] Among them, e k represents the classification error rate of the kth basic classifier in the Adaboost algorithm; x i Represents the training sample set T k-1 The i-th training sample in G k (x i ) represents the kth basic classifier for the training sample set T k-1 The classification result of the i-th training sample in y i Represents the training sample set T k-1 The actual classification result corresponding to the i-th training sample in P(G k (x i )≠y i ) indicates that the kth basic classifier in the Adaboost algorithm will train the sample set T k-1 The probability of misclassification of training samples in the training set T k-1 The total number of training samples misclassified by the kth basic classifier; w ki Represents the training sample set T during the kth training k-1 The weight value of the i-th training sample in . It should be noted that the number of training times for the credit risk rating assessment model based on the Adaboost algorithm is equal to the total number of classifiers of the basic classifier in the Adaboost algorithm. k (x i )≠y i ) is an indicator function, if G k )x i )≠y i If it is established, then I)G k )x i )≠y i )=1; if G k )x i )≠yi Not true, I(G k (x i )≠y i )=0.

[0059] Afterwards, according to the classification error rate of the kth basic classifier in the Adaboost algorithm, the weight of the kth basic classifier in the Adaboost algorithm is determined by the following formula:

[0060]

[0061] Among them, α k Represents the weight of the kth basic classifier in the Adaboost algorithm.

[0062] Afterwards, according to the weight of the kth basic classifier in the Adaboost algorithm, the training sample set T is updated by the following formula: k-1 The weight value distribution of the training sample set T k .

[0063]

[0064] Among them, D k (i) represents the training sample set T obtained after the kth training k The weight value distribution, z t represents the normalization constant,

[0065] After that, after determining the weight of each basic classifier in the Adaboost algorithm, a weighted voting mechanism is used to linearly combine the basic classifiers in the Adaboost algorithm to obtain the target evaluation model, namely:

[0066]

[0067] Where G(x) represents the target evaluation model; K represents the total number of classifiers in the basic classifier of the Adaboost algorithm; G k (x) represents the kth basic classifier.

[0068] S204: Obtain the credit data to be evaluated of the target object.

[0069] S205: Input the credit data to be evaluated into the target evaluation model to obtain the credit risk level corresponding to the target object.

[0070] The technical solution of the embodiment of the present invention is to obtain a sample credit data set; perform feature selection on the sample credit data set to obtain a key credit data set; construct a training sample set based on the key credit data set, and train a credit risk rating assessment model based on the Adaboost algorithm based on the training sample set to obtain a target assessment model; obtain the credit data to be assessed of the target object; input the credit data to be assessed into the target assessment model to obtain the credit risk level corresponding to the target object. The above technical solution, after obtaining the sample credit data set, performs a series of data preprocessing such as feature selection on the sample credit data set to obtain a training sample set, which reduces the amount of training data when training the credit risk rating assessment model and improves the data quality of training data when training the credit risk rating assessment model. Afterwards, the credit risk rating assessment model based on the Adaboost algorithm is trained according to the training sample set to obtain the target assessment model. On the one hand, the consumption of computing resources during the training of the credit risk rating assessment model is reduced, the model training time is reduced, and the data processing efficiency during the training of the credit risk rating assessment model is improved; on the other hand, the accuracy and reliability of the target assessment model are also improved. Afterwards, the target assessment model is used to process the target object's credit data to be assessed, so that the corresponding credit risk level of the target object is more accurate, thereby improving the accuracy of the user's credit risk assessment, reducing the credit risk in the credit business of banks and other financial institutions, allowing financial institutions to adjust credit limits and terms in a targeted manner, avoiding excessive concentration on high-risk projects, and thus reducing the overall risk of financial institutions' credit assets. At the same time, there is no need for credit approval personnel to review the borrower's loan application materials one by one, thereby improving the loan approval efficiency of banks and other financial institutions and shortening the borrower's business processing time.

[0071] Embodiment 3

[0072] Figure 3 This is a flowchart of a credit risk rating assessment method provided in Example 3 of the present invention. Based on the above example, this example further optimizes "constructing a training sample set based on a key credit data set" and provides an optional implementation plan. It should be noted that for the parts not described in detail in the example of the present invention, reference can be made to the relevant descriptions of other examples. Figure 3 As shown, the method includes:

[0073] S301. Obtain a sample credit data set.

[0074] S302: Perform feature selection on the sample credit data set to obtain a key credit data set.

[0075] S303. Based on the fuzzy C-means clustering algorithm, cluster the key credit data set to obtain a clustering result.

[0076] Among them, the fuzzy C-means algorithm (FCM) is a soft clustering method, and its core idea is to use membership to represent the relationship between data in a data set, so as to determine the data cluster to which each data belongs.

[0077] The clustering result refers to the result obtained after clustering the key credit data set; optionally, the clustering result includes information of each data cluster.

[0078] Specifically, the number of clusters can be determined according to the specific number of credit risk levels. For example, if the credit risk level is divided into the following 5 levels: normal, attention, substandard, suspicious and loss, the number of clusters is 5. After that, the number of clusters and the key credit data set are input into the fuzzy C-means clustering algorithm, and the clustering results are obtained after being processed by the fuzzy C-means clustering algorithm. The processing process of the fuzzy C-means clustering algorithm on the number of clusters and the key credit data set is as follows:

[0079] Step a: Initialize the fuzzy factor, current number of iterations, membership matrix, iteration allowable error and maximum number of iterations.

[0080] Specifically, the fuzzy factor, the current number of iterations, the membership matrix, the iteration allowable error and the maximum number of iterations may be initialized according to actual business requirements and the experience of those skilled in the art.

[0081] Step b: Check whether the current number of iterations is less than the maximum number of iterations.

[0082] Specifically, if the current number of iterations is less than the maximum number of iterations, then step c is performed; otherwise, step e is performed.

[0083] Step c: determining the cluster center according to the membership in the membership matrix, and updating the membership in the membership matrix according to the cluster center to obtain an updated membership matrix.

[0084] Specifically, according to the membership in the membership matrix, the cluster center is determined by the following cluster center determination formula:

[0085]

[0086] Among them, N j represents the jth cluster center; q represents the fuzzy factor, and the value of q is usually a real number greater than 1; u ij Indicates the membership degree of the i-th sample credit data in the key credit data set to the j-th cluster; x i represents the i-th sample credit data in the key credit dataset, and c represents the total number of sample credit data in the key credit dataset.

[0087] Afterwards, according to the cluster center, the membership in the membership matrix is ​​updated by the following membership update formula to obtain the updated membership matrix:

[0088]

[0089] Where n represents the number of clusters; N k represents the kth cluster center.

[0090] Step d: Check whether the objective function of the fuzzy C-means clustering algorithm satisfies the iteration stopping condition.

[0091] Among them, the objective function of the fuzzy C-means clustering algorithm is as follows:

[0092]

[0093] Among them, d ij Represents the i-th sample credit data x in the key credit data set i With the jth cluster center N j It should be noted that in order to minimize the objective function of the fuzzy C-means clustering algorithm, the objective function of the fuzzy C-means clustering algorithm needs to satisfy the following constraints:

[0094]

[0095] Where c represents the total number of sample credit data in the key credit dataset; n represents the number of clusters.

[0096] The iteration stop condition refers to the condition that stops the iteration of the fuzzy C-means clustering algorithm; optionally, the iteration stop condition can be determined according to the iteration allowable error and two adjacent objective functions. For example, the iteration stop condition can be: ||J t (U,V)-J t-1 (U,V)||<ε. Among them, J t (U, V) represents the objective function of the fuzzy C-means clustering algorithm at the tth iteration; J t-1 (U, V) represents the objective function of the fuzzy C-means clustering algorithm at the t-1th iteration.

[0097] Specifically, if the objective function of the fuzzy C-means clustering algorithm satisfies the iteration stop condition, step e is executed; otherwise, the current iteration number is increased by 1, and step b is continued.

[0098] Step e: output the clustering results.

[0099] It can be understood that clustering the key credit data set based on the fuzzy C-means clustering algorithm can bring together the sample credit data of the same credit risk level in the key credit data set, so as to better mine the potential information between the same type of data and provide more potential information for the subsequent multiple interpolation regression analysis.

[0100] Optionally, after obtaining the clustering results, the clustering labels of each data cluster in the clustering results can be one-hot encoded to obtain the clustering features corresponding to each data cluster; for each data cluster, the cluster corresponding to the data cluster is added to the feature information of each sample credit data in the data cluster to enrich the feature dimensions of the sample credit data in the key credit data set.

[0101] S304. Based on the clustering results, a chain equation multiple interpolation algorithm is used to perform data interpolation processing on the key credit data set to obtain a training sample set.

[0102] Among them, the chained equation multiple imputation algorithm (Multiple Imputation by Chained Equations, MICE) fills the missing values ​​in the data set through an iterative prediction model. It should be noted that in an embodiment of the present invention, the iterative prediction model used in the chained equation multiple imputation algorithm is an iterative prediction model based on the gradient boosting tree (Gradient Boosting Decision Tree, GBDT). Among them, GBDT is an integrated learning algorithm, and its core idea is to iteratively construct a series of decision trees, each decision tree attempts to correct the prediction error of the previous decision tree. In each round of iteration, for each training sample, the error of the current model is calculated; using the error of the current model as the target, a new decision tree is trained to obtain a new decision tree; in order to minimize the loss, the gradient of the loss function is estimated and the learning rate is determined; the result of multiplying the new decision tree by the learning rate is added to the current model to obtain an updated current model; repeat the above process until a predetermined number of iterations is reached, or until the performance of the model is no longer significantly improved, so as to obtain the final iterative prediction model.

[0103] For example, for training sample 1, in the first iteration, the predicted value of the current model for training sample 1 is calculated, and the difference between the actual observed value and the predicted value of training sample 1 is taken as the error of the current model, denoted as e i ; Use the error e of the current model i As the goal, train a new decision tree and get the following new decision tree:

[0104] y=a i +(b i *X)+e i ;

[0105] Among them, y represents the new decision tree, X represents the training sample 1, a i and b i is the preset coefficient.

[0106] Afterwards, the final iterative prediction model is obtained at the end of the iteration:

[0107]

[0108] Where n represents the predetermined number of iterations; e n Represents the error of the current model in the nth iteration.

[0109] Specifically, the minority cluster and majority cluster in the key credit data set can be determined based on the clustering results; the minority cluster is oversampled to obtain a new data cluster corresponding to the minority cluster; the chain equation multiple interpolation algorithm is used to interpolate the new data clusters corresponding to the majority cluster and the minority cluster to obtain a training sample set. Among them, the minority cluster refers to the data cluster with a small number of sample credit data; correspondingly, the majority cluster refers to the data cluster with a large number of sample credit data.

[0110] More specifically, the key credit data set can be classified according to the total data volume of each data cluster in the clustering results to obtain minority clusters and majority clusters. The total data volume refers to the total number of sample credit data in the data cluster. For example, if the clustering results include three data clusters, namely, data cluster 1, data cluster 2, and data cluster 3; wherein the total data volume of data cluster 1 is 50, the total data volume of data cluster 2 is 60, and the total data volume of data cluster 3 is 40, then the data cluster with the largest total data volume in the clustering results (i.e., data cluster 2) can be used as the majority cluster, and the remaining data clusters in the clustering results (i.e., data cluster 1 and data cluster 3) can be used as minority clusters.

[0111] Afterwards, an oversampling algorithm can be used to oversample the minority clusters to obtain new data clusters corresponding to the minority clusters. The oversampling algorithm can be preset according to actual business needs. For example, the oversampling algorithm can be one of a random oversampling algorithm, a synthetic minority oversampling algorithm (Synthetic Minority Oversampling Technique, SMOTE) and an adaptive synthetic sampling algorithm (Adaptive Synthetic Sampling, ADASYN). Based on the above example, the synthetic minority oversampling algorithm can be used to oversample data cluster 1 and data cluster 3 respectively to obtain new data clusters corresponding to data cluster 1 and new data clusters corresponding to data cluster 3.

[0112] Afterwards, the chain equation multiple interpolation algorithm is used to perform data interpolation processing on the new data clusters corresponding to the majority cluster and the minority cluster to obtain the training sample set.

[0113] It can be understood that according to the clustering results, the minority clusters and majority clusters in the key credit data set are determined, and the minority clusters in the key credit data set can be quickly screened out; then, the minority clusters are oversampled to obtain new data clusters corresponding to the minority clusters, which increases the amount of data in the minority clusters and solves the problem of data category imbalance in the key credit data set; then, the chain equation multiple interpolation algorithm is used to perform data interpolation on the new data clusters corresponding to the majority clusters and the minority clusters, respectively, to obtain the training sample set, which solves the problem of missing data in the key credit data set, thereby improving the data quality of the training sample set. In the process of solving the problem of missing data in the key credit data set, the problem of data category imbalance in the key credit data set is solved simultaneously, avoiding further expansion of the error of the interpolation process during the sampling process, thereby reducing the risk of overfitting in the model training process, improving data utilization, and thus improving the robustness of the target evaluation model in the face of different situations.

[0114] S305: Train the credit risk rating assessment model based on the Adaboost algorithm according to the training sample set to obtain a target assessment model.

[0115] S306: Obtain the credit data of the target object to be evaluated.

[0116] S307: Input the credit data to be evaluated into the target evaluation model to obtain the credit risk level corresponding to the target object.

[0117] The technical solution of the embodiment of the present invention is to obtain a sample credit data set; perform feature selection on the sample credit data set to obtain a key credit data set; perform clustering processing on the key credit data set based on the fuzzy C-means clustering algorithm to obtain a clustering result; according to the clustering result, use the chain equation multiple interpolation algorithm to perform data interpolation processing on the key credit data set to obtain a training sample set; train the credit risk rating assessment model based on the Adaboost algorithm according to the training sample set to obtain a target assessment model; obtain the credit data to be assessed of the target object; input the credit data to be assessed into the target assessment model to obtain the credit risk level corresponding to the target object. The above technical solution, after obtaining the sample credit data set, obtains the training sample set after a series of data preprocessing such as feature selection, fuzzy clustering processing, minority oversampling processing and multiple interpolation processing on the sample credit data set, solves the problem of data category imbalance and data missing in the training data when training the credit risk rating assessment model, reduces the data amount of the training data when training the credit risk rating assessment model, and thus improves the data quality of the training data when training the credit risk rating assessment model. In the process of solving the problem of missing data, the problem of data category imbalance is solved simultaneously, which avoids further expanding the error of the interpolation process during the sampling process, thereby reducing the risk of overfitting and improving data utilization. Afterwards, the credit risk rating assessment model based on the Adaboost algorithm is trained according to the training sample set to obtain the target assessment model. On the one hand, the consumption of computing resources during the training of the credit risk rating assessment model is reduced, the model training time is reduced, and the data processing efficiency during the training of the credit risk rating assessment model is improved; on the other hand, the accuracy and reliability of the target assessment model are also improved. Afterwards, the target assessment model is used to process the credit data to be assessed of the target object, so that the credit risk level corresponding to the target object is more accurate, thereby improving the accuracy of the credit risk assessment of the user, reducing the credit risk in the credit business of banks and other financial institutions, and allowing financial institutions to adjust the credit limit and term in a targeted manner to avoid excessive concentration on high-risk projects, thereby reducing the overall risk of the credit assets of financial institutions. At the same time, there is no need for credit approval personnel to review the loan application materials of borrowers one by one, thereby improving the loan approval efficiency of banks and other financial institutions and shortening the business processing time of borrowers.

[0118] Embodiment 4

[0119] Figure 4 This is a schematic diagram of the structure of a credit risk level assessment device provided by the fourth embodiment of the present invention. This embodiment can be applied to the situation of assessing the credit risk level of the customers of financial institutions. The device can be implemented in the form of hardware and / or software and can be configured in an electronic device. Figure 4 As shown, the device comprises:

[0120] The credit data acquisition module 401 is used to acquire the credit data of the target object.

[0121] The credit risk rating assessment module 402 is used to input the credit data to be assessed into the target assessment model to obtain the credit risk rating corresponding to the target object; wherein the target assessment model is obtained by training a credit risk rating assessment model based on the Adaboost algorithm based on a sample credit data set.

[0122] The technical solution of the embodiment of the present invention obtains the credit data to be evaluated of the target object; inputs the credit data to be evaluated into the target evaluation model to obtain the credit risk level corresponding to the target object; wherein the target evaluation model is obtained by training the credit risk level evaluation model based on the Adaboost algorithm based on the sample credit data set. The above technical solution processes the credit data to be evaluated of the target object through the target evaluation model, thereby obtaining the credit risk level corresponding to the target object, improving the accuracy of the credit risk assessment of the user and reducing the credit risk in the credit business of banks and other financial institutions. At the same time, there is no need for credit approval personnel to review the loan application materials of the borrower one by one, thereby improving the loan approval efficiency of banks and other financial institutions and shortening the business processing time of the borrower.

[0123] Optionally, the device further includes a target evaluation model determination module, wherein the target evaluation model determination module includes:

[0124] A sample credit data set acquisition unit, used to acquire a sample credit data set;

[0125] A key credit data set determination unit, used for performing feature selection on a sample credit data set to obtain a key credit data set;

[0126] The target assessment model determination unit is used to construct a training sample set according to the key credit data set, and train the credit risk rating assessment model based on the Adaboost algorithm according to the training sample set to obtain the target assessment model.

[0127] Optionally, a target assessment model determination unit includes:

[0128] A clustering result determination subunit is used to perform clustering processing on the key credit data set based on the fuzzy C-means clustering algorithm to obtain a clustering result;

[0129] The training sample set determines the subunits, which are used to perform data interpolation processing on the key credit data set based on the clustering results and adopt the chain equation multiple interpolation algorithm to obtain the training sample set.

[0130] Optionally, the training sample set determines a subunit, specifically for:

[0131] Based on the clustering results, determine the minority clusters and majority clusters in the key credit data set;

[0132] Oversampling the minority clusters to obtain new data clusters corresponding to the minority clusters;

[0133] The chain equation multiple interpolation algorithm is used to perform data interpolation on the new data clusters corresponding to the majority cluster and the minority cluster to obtain the training sample set.

[0134] Optionally, based on the clustering results, minority clusters and majority clusters in the key credit data set are determined, specifically: based on the total data volume of each data cluster in the clustering results, the key credit data set is classified to obtain minority clusters and majority clusters.

[0135] Optionally, a key credit data set determination unit is used to:

[0136] determining a first data feature and a second data feature in a sample credit data set;

[0137] Calculate the correlation between each sub-data feature in the second data feature and the first data feature based on the Pearson correlation coefficient;

[0138] According to the correlation between each sub-data feature in the second data feature and the first data feature, feature selection is performed on the sample credit data set to obtain a key credit data set.

[0139] The credit risk level assessment device provided in the embodiment of the present invention can execute the credit risk level assessment method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing each credit risk level assessment method.

[0140] According to an embodiment of the present invention, the present invention also provides an electronic device, a readable storage medium and a computer program product.

[0141] Embodiment 5

[0142] Figure 5 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0143] like Figure 5 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, ROM 12 and RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0144] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0145] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a credit risk rating assessment method.

[0146] In some embodiments, the credit risk rating assessment method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the credit risk rating assessment method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the credit risk rating assessment method in any other appropriate manner (e.g., by means of firmware).

[0147] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0148] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0149] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0150] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0151] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0152] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0153] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0154] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A credit risk level assessment method, characterized in that: include: Obtain the credit data of the target object to be evaluated; The credit data to be evaluated is input into a target evaluation model to obtain a credit risk level corresponding to the target object; wherein the target evaluation model is obtained by training a credit risk level evaluation model based on an Adaboost algorithm based on a sample credit data set.

2. The method according to claim 1, characterized in that The determination process of the target evaluation model is as follows: Get a sample credit dataset; Performing feature selection on the sample credit data set to obtain a key credit data set; A training sample set is constructed according to the key credit data set, and a credit risk rating assessment model based on the Adaboost algorithm is trained according to the training sample set to obtain a target assessment model.

3. The method according to claim 2, characterized in that A training sample set is constructed based on the key credit data set, including: Based on the fuzzy C-means clustering algorithm, clustering processing is performed on the key credit data set to obtain a clustering result; According to the clustering results, a chain equation multiple interpolation algorithm is used to perform data interpolation processing on the key credit data set to obtain a training sample set.

4. The method according to claim 3, characterized in that According to the clustering result, the chain equation multiple interpolation algorithm is used to perform data interpolation processing on the key credit data set to obtain a training sample set, including: Determining a minority cluster and a majority cluster in the key credit data set according to the clustering result; Oversampling the minority cluster to obtain a new data cluster corresponding to the minority cluster; A chain equation multiple interpolation algorithm is used to perform data interpolation processing on the new data clusters corresponding to the majority cluster and the minority cluster respectively, so as to obtain a training sample set.

5. The method according to claim 4, characterized in that Determining the minority cluster and the majority cluster in the key credit data set according to the clustering result includes: According to the total data volume of each data cluster in the clustering result, the key credit data set is classified to obtain minority clusters and majority clusters.

6. The method according to claim 2, characterized in that The step of performing feature selection on the sample credit data set to obtain a key credit data set includes: determining a first data feature and a second data feature in the sample credit data set; Calculate the correlation between each sub-data feature in the second data feature and the first data feature based on the Pearson correlation coefficient; According to the correlation between each sub-data feature in the second data feature and the first data feature, feature selection is performed on the sample credit data set to obtain a key credit data set.

7. A credit risk level assessment device, characterized in that: include: A module for acquiring credit data to be evaluated, used to acquire credit data to be evaluated of a target object; The credit risk rating assessment module is used to input the credit data to be assessed into a target assessment model to obtain the credit risk rating corresponding to the target object; wherein the target assessment model is obtained by training a credit risk rating assessment model based on the Adaboost algorithm based on a sample credit data set.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the credit risk rating assessment method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the credit risk rating assessment method according to any one of claims 1 to 6 when executed.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the credit risk rating assessment method according to any one of claims 1 to 6 is implemented.

Images