CN118037440A - Trusted data processing method and system for comprehensive credit system - Google Patents
Trusted data processing method and system for comprehensive credit system Download PDFInfo
- Publication number
- CN118037440A CN118037440A CN202410423023.7A CN202410423023A CN118037440A CN 118037440 A CN118037440 A CN 118037440A CN 202410423023 A CN202410423023 A CN 202410423023A CN 118037440 A CN118037440 A CN 118037440A
- Authority
- CN
- China
- Prior art keywords
- credit
- data
- applicant
- processing
- edge
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 22
- 230000006399 behavior Effects 0.000 claims abstract description 181
- 238000012545 processing Methods 0.000 claims abstract description 167
- 238000004088 simulation Methods 0.000 claims abstract description 91
- 238000000034 method Methods 0.000 claims abstract description 51
- 238000005457 optimization Methods 0.000 claims abstract description 37
- 238000012502 risk assessment Methods 0.000 claims abstract description 22
- 238000013135 deep learning Methods 0.000 claims abstract description 20
- 238000004458 analytical method Methods 0.000 claims description 60
- 238000012549 training Methods 0.000 claims description 58
- 230000004927 fusion Effects 0.000 claims description 34
- 230000009471 action Effects 0.000 claims description 32
- 238000013507 mapping Methods 0.000 claims description 25
- 238000013210 evaluation model Methods 0.000 claims description 19
- 238000012512 characterization method Methods 0.000 claims description 16
- 108090000623 proteins and genes Proteins 0.000 claims description 16
- 238000005070 sampling Methods 0.000 claims description 16
- 238000013475 authorization Methods 0.000 claims description 15
- 238000011161 development Methods 0.000 claims description 13
- 238000013433 optimization analysis Methods 0.000 claims description 11
- 238000007418 data mining Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000013139 quantization Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 238000012163 sequencing technique Methods 0.000 claims description 6
- 238000013075 data extraction Methods 0.000 claims description 5
- 238000012098 association analyses Methods 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 17
- 238000013480 data collection Methods 0.000 abstract description 5
- 230000010354 integration Effects 0.000 abstract description 4
- 238000011156 evaluation Methods 0.000 description 22
- 238000004422 calculation algorithm Methods 0.000 description 18
- 238000005516 engineering process Methods 0.000 description 14
- 238000010586 diagram Methods 0.000 description 11
- 238000010801 machine learning Methods 0.000 description 10
- 238000007619 statistical method Methods 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 7
- 230000000875 corresponding effect Effects 0.000 description 7
- 238000007726 management method Methods 0.000 description 7
- 238000005065 mining Methods 0.000 description 7
- 230000008901 benefit Effects 0.000 description 6
- 238000004140 cleaning Methods 0.000 description 5
- 238000009826 distribution Methods 0.000 description 5
- 230000004044 response Effects 0.000 description 5
- 238000012954 risk control Methods 0.000 description 4
- 238000012550 audit Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 125000004122 cyclic group Chemical group 0.000 description 3
- 238000000691 measurement method Methods 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 230000002787 reinforcement Effects 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000012300 Sequence Analysis Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000002596 correlated effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012407 engineering method Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000000342 Monte Carlo simulation Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013079 data visualisation Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011985 exploratory data analysis Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 239000002096 quantum dot Substances 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000010206 sensitivity analysis Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Landscapes
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to a trusted data processing method and system of a comprehensive credit system. The method comprises the following steps: acquiring a bank information center network interface, carrying out heterogeneous data collection cloud integration processing on the bank information center network interface, generating applicant credit cloud data, carrying out deep learning risk assessment processing based on the applicant credit cloud data, generating bank credit risk score data, carrying out intelligent credit product matching and meta space credit simulation processing on the bank credit risk score data, generating bank credit behavior prediction report data, carrying out dynamic credit decision optimization processing on the bank credit behavior prediction report data, and generating bank credit trust decision report data; the invention improves the decision transparency based on the dynamic decision of the prediction report, the applicant clearly knows the decision process, and utilizes the meta universe to simulate the future credit behaviors of the applicant and increase the future behavior prediction capability.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a trusted data processing method and system of a comprehensive credit system.
Background
The traditional bank credit method has limitation in facing large-scale data and complex decisions, so that the introduction of a machine learning technology becomes an innovative solution. Machine learning is a technology in the field of artificial intelligence, based on statistics and data analysis, automatic learning is performed from data by constructing and training a model, complex problems are processed through prediction and decision making, characteristics related to trust decision making are automatically selected and extracted from massive data by using a machine learning algorithm such as information gain, L1 regularization and the like, the data are converted into effective input which can be used for model training, hidden modes and group structures in the data are explored by using an unsupervised learning algorithm such as cluster analysis, association rule mining and the like, potential risk factors and behavior modes are found by helping banks, traditional bank trust methods are often based on artificial experience and rules, the artificial decision making is easily influenced by subjective factors, inconsistency and uncertainty of decision making results are caused, a large amount of complex data and multidimensional information cannot be effectively processed, and association and modes hidden behind the data are difficult to find.
Disclosure of Invention
The invention provides a credit authorization data processing method and system of a comprehensive credit system, which aim to solve at least one technical problem.
In order to achieve the above object, the present invention provides a method for processing credit data of an integrated credit system, the method comprising the steps of:
Step S1: acquiring a bank information center network interface, collecting heterogeneous data of the bank information center network interface, integrating the heterogeneous data in a cloud, and generating credit cloud data of the applicant;
Step S2: performing deep learning risk assessment processing based on the credit cloud data of the applicant to generate bank credit risk scoring data;
Step S3: performing intelligent credit product list matching processing on the bank credit risk scoring data to generate bank credit product matching list data;
Step S4: performing meta-universe credit simulation processing on the matching list data of the bank credit products to generate bank credit behavior prediction report data;
Step S5: and carrying out dynamic credit decision optimization processing on the bank credit behavior prediction report data to generate bank credit decision report data.
The invention provides a credit data processing method of a comprehensive credit system, which is characterized in that heterogeneous data source acquisition processing is carried out on a network interface of a bank information center to generate a multi-source information set of an applicant, heterogeneous data collection can acquire data from different systems and channels, credit cloud data of the applicant has diversified information sources to reflect credit conditions of the applicant more comprehensively, association relations among different data can be established by integrating the heterogeneous data, potential relations and interaction among the data are revealed, more accurate credit evaluation and prediction capability are provided, deep learning risk evaluation processing is carried out based on the credit data of the applicant to generate bank credit risk scoring data, real-time evaluation and feedback of the credit conditions of the applicant are realized through the use of the cloud data, the bank can immediately acquire latest credit information, timely and accurate risk evaluation is carried out, the deep learning risk evaluation can automatically process a large amount of cloud data, the requirement of manual intervention is reduced, and the efficiency of risk evaluation is improved. The bank can rapidly process credit applications and reduce risks of errors and loopholes, intelligent credit product matching is conducted on bank credit risk scoring data, bank credit product matching list data are generated, intelligent credit product matching can conduct personalized matching according to credit risk scoring data of borrowers, specific conditions of each borrower are guaranteed, each borrower is guaranteed to obtain credit products which are most suitable for risk bearing capacity and repayment capacity, matching degree and adaptability of the products are improved, and benefits of the products can be maximized by the bank through intelligent credit product matching. The matching list data can accurately determine key information such as credit line, interest rate, repayment limit and the like required by borrowers, the maximum lending benefit and risk control are realized, the intelligent credit product matching can reduce credit risk, the matching list data can accurately evaluate the risk level of the borrowers by analyzing credit risk scoring data, and accordingly, the proper credit product is recommended, the offence risk can be reduced, the repayment rate is improved, and finally, the credit loss of a bank is reduced, the bank credit product matching list data is subjected to meta-universe credit simulation processing, the bank credit behavior prediction report data is generated, the credit behaviors of the borrowers can be predicted more accurately through meta-universe credit simulation processing, misjudgment and misleading are avoided, the accuracy of prediction results is improved, the meta-universe credit simulation processing has the modeling capability of historical and future data, time and space can be spanned, the cross-space credit behavior prediction is provided, the bank is helped to make more distant decisions, the meta-universe credit simulation processing can provide novel decision basis and data insight, the bank is helped to find association and modes behind the data decisions, the promotion innovation and the credit product matching list data are subjected to meta-universe credit prediction report data, the current decision report data is optimized, the real-time error is better, the reliability of the credit prediction report data is optimized, and the current market error is better, and the reliability of the reliability is better, and the reliability is better.
In one embodiment of the present specification, there is provided a credit authorization data processing system for an integrated credit system, comprising:
at least one processor;
a memory communicatively coupled to the at least one processor;
The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the credit data processing method of the integrated credit system as recited in any one of the preceding claims.
The invention provides a credit data processing system of a comprehensive credit system, which can realize the credit data processing method of any comprehensive credit system, obtain a bank information center network interface, perform heterogeneous data collection cloud integration processing on the bank information center network interface, generate applicant credit cloud data, perform deep learning risk assessment processing and intelligent credit product matching based on the applicant credit cloud data, generate bank credit product matching list data, perform meta-universe credit simulation processing on the bank credit product matching list data, generate bank credit behavior prediction report data, perform dynamic credit decision optimization processing on the bank credit behavior prediction report data, generate bank credit decision report data, and the system internally follows a set instruction set to complete the operation steps of the method so as to promote completion of the credit data processing method of the comprehensive credit system.
The invention combines a multidisciplinary and multisypic model to provide a credit data processing method of a comprehensive credit system, helps banks to find potential risks and behavior modes in the credit process, finds out relations and modes hidden behind data, promotes innovation decision and business development, simultaneously realizes real-time evaluation and feedback of credit conditions of the applicant, timely makes accurate risk evaluation, rapidly processes credit applications, and reduces risks of errors and loopholes.
Drawings
FIG. 1 is a flow chart of steps of a method for processing credit data of a comprehensive credit system according to the present invention;
FIG. 2 is a detailed implementation step flow diagram of step S1;
FIG. 3 is a detailed implementation step flow diagram of step S2;
FIG. 4 is a detailed implementation step flow diagram of step S3;
FIG. 5 is a detailed implementation step flow diagram of step S4;
fig. 6 is a detailed implementation step flow diagram of step S5.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The embodiment of the application provides a credit authorization data processing method and system of a comprehensive credit system. The execution main body of the credit authorization data processing method and the system of the integrated credit system comprises, but is not limited to, the system is carried: mechanical devices, data processing platforms, cloud server nodes, network transmission devices, etc. may be considered general purpose computing nodes of the present application. The data processing platform includes, but is not limited to: at least one of an audio management system, an image management system and an information management system.
Referring to fig. 1 to 6, the present invention provides a method for processing credit authorization data of an integrated credit system, the method comprising the following steps:
Step S1: acquiring a bank information center network interface, collecting heterogeneous data of the bank information center network interface, integrating the heterogeneous data in a cloud, and generating credit cloud data of the applicant;
Step S2: performing deep learning risk assessment processing based on the credit cloud data of the applicant to generate bank credit risk scoring data;
Step S3: performing intelligent credit product list matching processing on the bank credit risk scoring data to generate bank credit product matching list data;
Step S4: performing meta-universe credit simulation processing on the matching list data of the bank credit products to generate bank credit behavior prediction report data;
Step S5: and carrying out dynamic credit decision optimization processing on the bank credit behavior prediction report data to generate bank credit decision report data.
The invention provides a credit data processing method of a comprehensive credit system, which is characterized in that heterogeneous data source acquisition processing is carried out on a network interface of a bank information center to generate a multi-source information set of an applicant, heterogeneous data collection can acquire data from different systems and channels, credit cloud data of the applicant has diversified information sources to reflect credit conditions of the applicant more comprehensively, association relations among different data can be established by integrating the heterogeneous data, potential relations and interaction among the data are revealed, more accurate credit evaluation and prediction capability are provided, deep learning risk evaluation processing is carried out based on the credit data of the applicant to generate bank credit risk scoring data, real-time evaluation and feedback of the credit conditions of the applicant are realized through the use of the cloud data, the bank can immediately acquire latest credit information, timely and accurate risk evaluation is carried out, the deep learning risk evaluation can automatically process a large amount of cloud data, the requirement of manual intervention is reduced, and the efficiency of risk evaluation is improved. The bank can rapidly process credit applications and reduce risks of errors and loopholes, intelligent credit product matching is conducted on bank credit risk scoring data, bank credit product matching list data are generated, intelligent credit product matching can conduct personalized matching according to credit risk scoring data of borrowers, specific conditions of each borrower are guaranteed, each borrower is guaranteed to obtain credit products which are most suitable for risk bearing capacity and repayment capacity, matching degree and adaptability of the products are improved, and benefits of the products can be maximized by the bank through intelligent credit product matching. The matching list data can accurately determine key information such as credit limit, interest rate, repayment limit and the like required by borrowers, the maximum lending benefit and risk control are realized, the intelligent credit product matching can reduce credit risk, the matching list data can accurately evaluate the risk level of the borrowers by analyzing credit risk scoring data, and accordingly, the proper credit product is recommended, the offence risk can be reduced, the repayment rate is improved, finally, the credit loss of a bank is reduced, the bank credit product matching list data is subjected to meta-universe credit simulation processing, the bank credit behavior prediction report data is generated, credit behaviors of the borrowers can be predicted more accurately through meta-universe credit simulation processing, misjudgment and misleading are avoided, the accuracy of prediction results is improved, the meta-universe credit simulation processing has the modeling capability of historical and future data, time and space crossing can be provided, the cross-space credit behavior prediction is facilitated, the bank is helped to make a more distant decision, the meta-universe credit simulation processing can provide novel decision basis and data insight, the bank is helped to find a potential decision trend and a decision, the bank is pushed to innovated, the business behavior prediction and business prediction data are better, the credit response is better optimized, the credit response condition of the credit response data is better, and the current credit response state is better, and the credit response state is better reliability is generated.
In the embodiment of the present invention, as described with reference to fig. 1, the step flow diagram of a credit authorization data processing method of a comprehensive credit system of the present invention is provided, in this example, the credit authorization data processing method of the comprehensive credit system includes the following steps:
Step S1: acquiring a bank information center network interface, collecting heterogeneous data of the bank information center network interface, integrating the heterogeneous data in a cloud, and generating credit cloud data of the applicant;
In the embodiment of the invention, heterogeneous data source acquisition processing is performed on a bank information center network interface, an applicant multi-source information set is generated, cloud big data processing is performed on the basis of the applicant multi-source information set, applicant cloud information portrait data is generated, credit evaluation model creation processing is performed on the applicant cloud information portrait data, applicant credit evaluation model data is generated, and credit cloud data generation processing is performed on the applicant credit evaluation model data, so that applicant credit cloud data is generated.
Step S2: performing deep learning risk assessment processing based on the credit cloud data of the applicant to generate bank credit risk scoring data;
In the embodiment of the invention, data mining and extraction processing is carried out on the credit cloud data of the applicant to generate the basic information data of the applicant, training processing is carried out on the basic information data of the applicant by utilizing deep learning edge training processing to generate an applicant credit edge training data set, reasoning optimization analysis is carried out on the basis of the applicant credit edge training data set to generate applicant credit reasoning data, risk assessment processing is carried out on the applicant credit reasoning data to generate bank credit risk scoring data, credit score storage management processing is carried out on the bank credit risk scoring data to generate bank credit risk scoring data.
Step S3: performing intelligent credit product list matching processing on the bank credit risk scoring data to generate bank credit product matching list data;
In the embodiment of the invention, credit risk element decoding processing is carried out on credit risk scoring data of a bank to generate credit risk element vector data, product characteristic quantization processing is utilized based on the credit risk element vector data to generate credit product characteristic quantum data, corresponding fusion matching algorithm (such as similarity calculation or Euclidean distance measurement method) is utilized to carry out fusion matching calculation on the credit product characteristic quantum data to generate credit product fusion matching degree data, matching degree quantum rank ordering processing is carried out on the credit product fusion matching degree data to generate credit product matching measurement sub-ordering data, credit product intelligent list generation is carried out on the credit product matching degree quantum ordering data to generate bank credit product matching list data.
Step S4: performing meta-universe credit simulation processing on the matching list data of the bank credit products to generate bank credit behavior prediction report data;
In the embodiment of the invention, meta-universe credit view modeling processing is carried out on bank credit product matching list data to generate super credit view characterization data, heterogeneous character modeling processing is carried out on the basis of the super credit view characterization data to generate heterogeneous character credit gene data, heterogeneous character credit behavior simulation processing is carried out on the heterogeneous character credit gene data to generate heterogeneous character credit behavior simulation field data, credit behavior pattern pedigree fusion analysis is carried out on the heterogeneous character credit behavior simulation field data to generate fusion credit behavior pattern pedigree data, credit behavior prediction report generation processing is carried out on the fusion credit behavior pattern pedigree data to generate bank credit behavior prediction report data.
Step S5: and carrying out dynamic credit decision optimization processing on the bank credit behavior prediction report data to generate bank credit decision report data.
In the embodiment of the invention, credit behavior prediction model construction processing is carried out on the credit behavior prediction report data of the bank to generate credit prediction model data, dynamic credit decision optimization processing is utilized to optimize the credit prediction model data to generate credit optimization decision report data, risk assessment strategy analysis processing is carried out on the credit optimization decision report data to generate bank credit decision report data.
Preferably, the specific steps of step S1 are:
Step S11: heterogeneous data source acquisition processing is carried out on a network interface of a bank information center, so that an applicant multi-source information set is generated;
step S12: carrying out cloud big data processing based on the applicant multi-source information set to generate applicant cloud information portrait data;
Step S13: carrying out credit evaluation model creation processing on the cloud information portrait data of the applicant to generate credit evaluation model data of the applicant;
Step S14: and carrying out credit cloud data generation processing on the credit evaluation model data of the applicant to generate the credit cloud data of the applicant.
The invention generates a multi-source information set by collecting information from different data sources by heterogeneous data source collection processing on a bank information center network interface, realizes the whole coverage and breadth expansion of the information of the applicant, forms a multi-source information set by collecting and integrating information of a plurality of data sources, provides comprehensive and rich data view angles, can build accurate and comprehensive applicant information portraits, compensates the problems such as data deficiency, deviation or incompleteness and the like possibly existing in a single data source to a certain extent, improves the quality and reliability of data, carries out cloud big data processing on the basis of the multi-source information set of the applicant, generates cloud big data processing technology, can efficiently process and analyze large-scale applicant information, can correlate and mine the multi-source information of the applicant through cloud big data processing, finds out the correlation relation and potential modes among different information, reveals deeper applicant characteristics and behavior modes, creates credit evaluation model data for the cloud information portraits, generates credit evaluation model data, accurately makes a decision-making decision making method, can accurately identify all the applicant's, can accurately and has high risk assessment decision making and decision making method, can quickly estimate the credit model data based on the credit risk factor, can be evaluated by a bank, and has high-risk factor and high-decision, and can quickly estimate the risk factor and can be evaluated by a bank decision-making factor, which can quickly and estimate the risk factor and estimate credit factor based on the credit, the credit cloud data of the applicant is generated, and the credit cloud data generation and processing can be combined with technical means such as data encryption and authority control, so that the safety and privacy protection of the credit data are ensured, and the leakage and abuse of sensitive information are avoided.
As an example of the present invention, referring to fig. 2, a detailed implementation step flow diagram of step S1 in fig. 1 is shown, where step S1 includes:
Step S11: heterogeneous data source acquisition processing is carried out on a network interface of a bank information center, so that an applicant multi-source information set is generated;
In the embodiment of the invention, data are acquired from a network interface of a bank information center, the isomerism of data sources is ensured, the data comprise data with different formats, different structures and different sources, the data acquisition processing is carried out on different data sources in the network interface of the bank information center, the operations comprising data extraction, data conversion, data integration and the like are carried out, so that the data sources with isomerism are acquired, and the data acquired from the isomerism data sources are integrated and combined to generate the applicant multisource information set. This information set contains applicant-related information from various data sources, such as personal basic information, financial information, lending history, etc.
Step S12: carrying out cloud big data processing based on the applicant multi-source information set to generate applicant cloud information portrait data;
in the embodiment of the invention, the multi-source information set of the applicant is processed and analyzed by utilizing cloud computing and big data processing technology. The method comprises the operations of data cleaning, data conversion, feature extraction, data mining and the like, so as to process large-scale data and extract useful information, and cloud information portrait data of an applicant is generated based on the cloud large-data processing result, wherein the information portrait data comprises comprehensive description of aspects of characteristics, behavior patterns, consumption preferences and the like of the applicant and is used for further credit assessment and modeling.
Step S13: carrying out credit evaluation model creation processing on the cloud information portrait data of the applicant to generate credit evaluation model data of the applicant;
According to the embodiment of the invention, the feature selection is carried out according to each feature in cloud information portrait data, the data cleaning, the data normalization and the missing value processing pretreatment operation are carried out on the cloud information portrait data so as to ensure the accuracy and the consistency of the data, a support vector machine model is selected for modeling according to the characteristics of credit evaluation requirements and problems, the cloud information portrait data is used as a training set, a credit evaluation model is trained by using a selected model algorithm, the super parameters of the model are adjusted and optimized according to the performance of a verification set in the training process, the trained model is evaluated by using a test set, and evaluation indexes of the model such as accuracy, precision, recall rate, F1 value and the like are calculated, so that the credit evaluation model data of the applicant is generated.
Step S14: and carrying out credit cloud data generation processing on the credit evaluation model data of the applicant to generate the credit cloud data of the applicant.
In the embodiment of the invention, a multi-source information set and credit evaluation model data of an applicant are input into a credit evaluation model, the credit of the applicant is evaluated and calculated by using the model, the model quantifies and predicts the credit of the applicant according to the multi-source information and model parameters of the applicant, the calculation of credit cloud data is performed according to the output result of the credit evaluation model, indexes such as credit score, risk level and credibility of the applicant are calculated, and the calculated credit cloud data is integrated with relevant data such as personal information and financial data of the applicant to generate the credit cloud data of the applicant.
Preferably, the specific steps of step S2 are:
step S21: performing data mining and extraction processing on the credit cloud data of the applicant to generate basic information data of the applicant;
step S22: training the basic information data of the applicant by using a deep learning edge training model to generate an applicant trust edge training data set;
Step S23: performing reasoning optimization analysis based on the applicant trust edge training data set to generate applicant trust reasoning data;
Step S24: and carrying out risk assessment processing on the credit reasoning data of the applicant to generate bank credit risk scoring data.
The invention generates the basic information data of the applicant by carrying out data mining and extracting processing on the credit cloud data of the applicant, can acquire comprehensive and detailed basic information of the applicant by carrying out mining and extracting processing on the credit cloud data of the applicant, the rich performance of the information comprehensively knows the credit condition and risk characteristics of the applicant, the scattered credit cloud data of the applicant is integrated into structured basic information data of the applicant by carrying out mining and extracting processing, the data is easier to manage, analyze and apply, the basic information data of the applicant is trained by utilizing deep learning edge training processing, an applicant credit edge training data set is generated, the basic information data of the applicant is processed by utilizing deep learning edge training processing, the learning capacity of a model is improved, the complicated nonlinear relation and mode can be automatically learned from the data, the prediction accuracy and the robustness of the model are improved, and then, the step involves carrying out reasoning optimization analysis on the credit edge decision credit condition of the applicant by utilizing various algorithms and models to simulate the credit edge decision situation of the applicant, the credit edge decision state, the action, the reward and the like are analyzed, and the future credit edge decision can be accurately and deeply estimated, and the future credit can be accurately estimated. The bank can accurately evaluate the credit condition and repayment capability of the applicant, reduce bad liabilities and default risks, store and manage the credit scores of the bank credit risk score data, generate the bank credit risk score data, and store and manage the bank credit risk score data so that the bank can monitor the credit risk change in real time and timely send out early warning signals, and can quickly respond to the risk condition, take necessary risk control measures and avoid potential loss.
As an example of the present invention, referring to fig. 3, a detailed implementation step flow diagram of step S2 in fig. 1 is shown, where step S2 includes:
step S21: performing data mining and extraction processing on the credit cloud data of the applicant to generate basic information data of the applicant;
in the embodiment of the invention, data exploratory analysis, including data visualization, statistical description, correlation analysis and the like, is performed on the credit cloud data of the applicant so as to know the characteristics and distribution conditions of the data, an association rule mining method is used for mining and extracting the original data, and characteristic engineering, including characteristic selection, characteristic construction, characteristic conversion and the like, is performed on personal information data so as to extract the characteristics with representative and predictive capabilities, and a cross verification method is used for optimizing and optimizing the model so as to generate the basic information data of the applicant.
Step S22: training the basic information data of the applicant by using a deep learning edge training model to generate an applicant trust edge training data set;
In the embodiment of the invention, an edge training model is constructed by utilizing a convolutional neural network algorithm, applicant basic information data is used as a training sample and is input into the edge training model, in the training process, the model is trained by an optimization algorithm (gradient descent) so as to improve the accuracy and the prediction capability of the model, and in the iterative training process, the model can better capture the relation between the applicant basic information and the trust edges by continuously adjusting model parameters and the optimization algorithm, so that an applicant trust edge training data set is generated;
Step S23: performing reasoning optimization analysis based on the applicant trust edge training data set to generate applicant trust reasoning data;
In the embodiment of the invention, firstly, the edge training data set containing various applicant credit cases is collected and subjected to data cleaning and preprocessing to remove abnormal values and missing data, the data quality is ensured, the edge training data set is subjected to feature extraction and analysis by using statistical analysis, machine learning or deep learning (such as convolutional neural network or cyclic neural network) and other technologies to identify the features related to credit edges, including personal information, financial condition, credit record and other features, meanwhile, the relevant features of credit edge situation are determined according to the extracted feature data, and the features are extracted and converted by utilizing a data mining or feature engineering method so as to better describe the credit edge situation, the feature point sampling is carried out on the credit edge situation feature information obtained by analysis by a corresponding feature point sampling method or strategy, to select representative feature sample points and ensure that the sampling result can fully reflect the feature distribution and condition of the original data, secondly, constructing a simulation space of the trust edge situation by using a space simulation method according to the feature point set obtained by sampling, extracting features related to decision states (such as financial conditions, credit records and the like of the applicant), actions (such as approval, refusal or additional audit and the like) and rewards (such as profits, risks and the like) from the trust edge training data set of the applicant, analyzing and modeling the features by utilizing statistical analysis or machine learning technology to identify modes and trends under different trust edge decisions, analyzing and synthesizing the obtained decision states, actions and rewards data, the method comprises the steps of establishing a mapping model among decision actions and rewards under different decision states, mapping and analyzing specific conditions of different trust edge decision strategies, then correlating the analyzed trust edge decision mapping strategies with the trust edge situation simulation space of the applicant so as to comprehensively consider decision influence conditions under different trust edge situations, and optimizing and analyzing the decision mapping strategies by utilizing an optimizing algorithm or an reasoning model so as to find an optimal decision scheme, wherein the optimal decision scheme comprises optimized decision results and recommended decision strategies, thereby accurately simulating trust decision processes under different trust edge decision situations, improving the efficiency and the accuracy of trust edge decision and finally generating trust reasoning data of the applicant.
Step S24: and carrying out risk assessment processing on the credit reasoning data of the applicant to generate bank credit risk scoring data.
In the embodiment of the invention, a credit scoring model is used, credit risk of the applicant is evaluated based on the credit reasoning data set of the applicant, a statistical analysis method is used for combining the credit reasoning data and the historical data of the applicant, risk indexes of the applicant are calculated, and the calculated risk indexes are converted into bank credit risk scoring data according to the risk evaluation standard of a bank.
Preferably, the specific steps of step S23 are:
Step S231: performing trust edge feature analysis on the trust edge training data set of the applicant to obtain trust edge feature data of the applicant;
In the embodiment of the invention, firstly, an edge training data set containing various applicant credit cases is collected and subjected to data cleaning and preprocessing to remove abnormal values and missing data, so that the data quality is ensured, and then, the edge training data set is subjected to feature extraction and analysis by using technologies such as statistical analysis, machine learning or deep learning (such as convolutional neural network or cyclic neural network) so as to identify features related to credit edges, including features such as personal information, financial conditions and credit record, and finally, the applicant credit edge feature data is obtained.
Step S232: performing trust edge situation feature extraction processing on the trust edge feature data of the applicant to obtain trust edge situation feature information data of the applicant;
In the embodiment of the invention, firstly, relevant characteristics of the credit-giving edge situation are determined according to the credit-giving edge characteristic data of the applicant, and the characteristics are extracted and converted by utilizing a data mining or characteristic engineering method so as to better describe the credit-giving edge situation, then, the extracted characteristic data are integrated, and the characteristic data are associated with different credit situations, such as financial conditions, behavior habits and the like of the applicant, so that the credit-giving edge situation characteristic information data of the applicant are finally obtained.
Step S233: performing edge situation feature point set sampling processing on the applicant trusted edge situation feature information data to obtain an applicant trusted edge situation feature point set;
in the embodiment of the invention, firstly, a sampling method and a sampling strategy, such as random sampling or layered sampling, are determined, and secondly, representative characteristic sample points are selected in the trusted edge situation characteristic information data set according to the sampling strategy, so that subsequent analysis and simulation can be more effectively carried out, the sampling result can be ensured to fully reflect the characteristic distribution and condition of the original data, and finally, the trusted edge situation characteristic point set of the applicant is obtained.
Step S234: performing edge situation simulation on the applicant trust edge situation feature point set to obtain an applicant trust edge situation simulation space;
In the embodiment of the invention, firstly, a simulation model of the credit giving edge situation is constructed by using a space simulation method according to the sampled feature point set, parameters and conditions of simulation such as time range, scene setting and the like are determined, then, simulation operation verification is carried out on the applicant credit giving edge situation feature point set to simulate the behaviors and decisions of the applicant under different credit situations, and the credit performances of the applicant under different situations can be observed and analyzed under control conditions, so that the applicant credit giving edge situation simulation space is finally obtained.
Step S235: and carrying out reasoning optimization analysis on the applicant trust edge training data set based on the applicant trust edge situation simulation space to generate applicant trust reasoning data.
In the embodiment of the invention, firstly, characteristics related to decision states (such as financial conditions, credit records and the like of the applicant), actions (such as approval, rejection or additional audits and the like) and rewards (such as profits, risks and the like) are extracted from an applicant trust edge training data set, secondly, the characteristics are analyzed and modeled by utilizing a statistical analysis or machine learning technology to identify modes and trends under different trust edge decisions, and the obtained decision states, actions and reward data are analyzed and integrated to reveal mapping relations between the decision actions and rewards under the different decision states, including establishing a mapping model between the decision states, the actions and the rewards, so as to map and analyze specific situations of different trust edge decision strategies, then, the analyzed trust edge decision mapping strategy is correlated with a trust edge situation simulation space of the applicant, so that decision influence situations under different trust edge situations are comprehensively considered, and the decision mapping strategy is optimized by utilizing an optimization algorithm or an inference model so as to find an optimal decision scheme, the trust edge decision result and the recommended trust edge decision can be accurately simulated, and the trust edge data of the applicant can be generated accurately.
The invention aims at identifying and extracting the edge characteristics related to trust by carrying out characteristic analysis on the trust edge training data set of the applicant, wherein the edge characteristics possibly relate to personal information, financial conditions, credit records and the like, and the credit condition of the applicant can be better understood by analyzing the characteristics. The key of the step is to be able to determine the key features of the applicant related to trust, identify the association between features and explore the potential non-linear relationship, thereby providing a richer and accurate data basis for subsequent credit assessment. Secondly, by performing a trust edge context feature extraction process on the trust edge feature data of the applicant, the purpose of this step is to associate the feature data with different credit contexts, such as financial conditions, behavior habits, etc. of the applicant. By associating the characteristic data with a context, the factors behind the applicant's credit behavior and decisions can be better understood. In particular, the key to this step is the ability to map feature data into different context spaces and identify changes and effects of features in different contexts, thereby providing a more detailed and comprehensive data base for subsequent credit assessment. Then, by performing an edge context feature point set sampling process on applicant trusted edge context feature information data, the purpose of this step is to select a representative sample point set from a large number of context feature data for more efficient subsequent analysis and simulation. By sampling, the dimension and complexity of the data can be reduced, and the efficiency and feasibility of subsequent processing can be improved. In particular, by selecting an appropriate sampling method and sample number, it is ensured that the set of sampling points can adequately cover various credit scenarios, and is representative and interpretable. Next, by performing an edge context simulation on the set of applicant's trusted edge context feature points, the purpose of this step is to build a simulation model of the trusted edge context based on the real data to simulate the applicant's behavior and decisions in different credit contexts. Through simulation, the credit performance of the applicant under different situations can be observed and analyzed under control conditions, so that basic data guarantee is provided for the subsequent assessment of the credit risk. By establishing mathematical description and parameter setting of a simulation space, a simulation experiment is carried out and simulation data of a trust edge situation is generated, so that a richer and reliable data basis is provided for subsequent credit evaluation. Finally, by carrying out inference optimization analysis on the training data set of the trust edge of the applicant based on the simulation space of the trust edge situation of the applicant, the step aims to strengthen and optimize the training data set through the simulation data, so that potential problems in the training data set can be found and corrected, trust inference data with more representativeness and reliability can be generated, and a more accurate and reliable basis is provided for final credit evaluation.
Preferably, the specific steps of step S235 are:
performing trust edge decision state analysis on the trust edge training data set of the applicant to obtain trust edge decision state data of the applicant;
In the embodiment of the invention, firstly, the characteristics related to the decision states, such as financial conditions, credit records and the like of the applicant, are extracted from the applicant trust edge training data set, secondly, the characteristics are analyzed and modeled by utilizing a statistical analysis or machine learning technology to identify modes and trends under different decision states, and then, the decision states of each applicant under different situations are described according to analysis results, including passing, rejecting or needing further examination and the like, so that the trust edge decision state data of the applicant is finally obtained.
Preferably, performing trust edge decision action analysis on the trust edge training data set of the applicant to obtain trust edge decision action data of the applicant;
In the embodiment of the invention, firstly, information related to decision actions, such as approval, rejection or additional audit, is extracted from an applicant trust edge training data set, secondly, decision action frequencies and modes under different situations are analyzed to know the preference and decision flow of a decision maker, and then, the decision action situation taken by each applicant under different situations is described according to the analysis result, so that applicant trust edge decision action data is finally obtained.
Preferably, carrying out credit edge decision rewarding analysis on the credit edge training data set of the applicant to obtain credit edge decision rewarding data of the applicant;
In the embodiment of the invention, firstly, information related to decision results is extracted from the trust edge training data set of the applicant, such as loan amount, interest rate and the like, and secondly, rewarding conditions under different decision results, including profit, risk and the like, are analyzed, and then, the rewarding conditions obtained by each applicant under different decision results are described according to the analysis results, so that the trust edge decision rewarding data of the applicant is finally obtained.
Preferably, performing trust edge decision mapping analysis on the trust edge decision state data, the trust edge decision action data and the trust edge decision rewarding data of the applicant to obtain trust edge decision mapping strategy data of the applicant;
in the embodiment of the invention, firstly, the obtained decision state, action and rewarding data are analyzed and synthesized to reveal the mapping relation between the decision action and rewarding under different decision states, including establishing a mapping model between the decision state, action and rewarding, so as to map, analyze and determine specific conditions of different trust edge decision strategies, and finally obtain the trust edge decision mapping strategy data of the applicant.
Preferably, the applicant trust edge decision mapping strategy data is subjected to reasoning optimization analysis based on the applicant trust edge situation simulation space, so that applicant trust reasoning data is generated.
In the embodiment of the invention, firstly, the decision-making mapping strategy of the trust edge of the applicant obtained by analysis is correlated with the simulation space of the trust edge situation simulation of the applicant so as to comprehensively consider the decision influence situations under different trust edge situations, and secondly, the decision-making mapping strategy is optimally analyzed by utilizing an optimization algorithm or an inference model so as to find an optimal decision scheme, comprising an optimized decision result and a recommended decision strategy, thereby accurately simulating trust decision processes under different trust edge decision situations, improving the efficiency and the accuracy of trust edge decision and finally generating trust inference data of the applicant.
The invention firstly carries out credit edge decision state analysis on the credit edge training data set of the applicant so as to identify and extract information related to the credit decision state, wherein the information comprises the current credit state of the applicant, such as whether the application passes, is refused or is in a pending state, so that the relevance between different credit edge decision states can be determined, the factors causing different states are analyzed, and possible modes and trends are identified, thereby providing deeper understanding and reference for bank decisions. Secondly, by carrying out trust edge decision action analysis on the trust edge training data set of the applicant so as to extract action information related to trust decisions, the actions comprise actions or operations of the applicant, such as submitting data, filling in application forms, taking investigation and the like, and more specific data support can be provided for solving the credit actions of the applicant by identifying and analyzing the frequency, sequence and influencing factors of different decision actions. Then, by performing a trust edge decision reward analysis on the applicant's trust edge training data set to extract rewards information related to trust decisions, including benefits or rewards that the applicant obtains from their behavior or decisions, it is also possible to identify and analyze therefrom the types, sizes and distribution of different decision rewards, as well as incentive mechanisms and potential influencing factors related to rewards. Next, by performing trust edge decision mapping analysis on the trust edge decision state data, the trust edge decision action data and the trust edge decision rewarding data of the applicant, the obtained decision state, action and rewarding data can be analyzed and synthesized to reveal the mapping relation between the decision action and rewards under different decision states, including establishing a mapping model between the decision state, the action and the rewards, so as to map and analyze the advantages and disadvantages of different trust edge decision strategies, and propose optimization suggestions to improve the effect and the result of the decision strategy. Finally, by carrying out reasoning optimization analysis on the applicant trust edge decision mapping strategy data based on the applicant trust edge situation simulation space, the decision mapping strategy can be verified and optimized by utilizing the simulation data, and the performances and the stability of different trust edge decision strategies can be evaluated, so that more reliable and effective trust reasoning data can be generated, and more accurate basis is provided for credit evaluation and decision.
Preferably, the specific steps of step S3 are:
step S31: carrying out credit risk element decoding processing on the credit risk scoring data of the bank to generate credit risk element vector data;
step S32: generating credit product characteristic quantum data by utilizing product characteristic quantization processing based on credit risk element vector data;
Step S33: carrying out fusion matching calculation on credit product characteristic quantum data to generate credit product fusion matching degree data;
Step S34: carrying out matching degree quantum-level sequencing treatment on the credit product fusion matching degree data to generate credit product matching degree quantum sequencing data;
step S35: and carrying out credit product intelligent list matching processing on the credit product matching degree quantum sorting data to generate bank credit product matching list data.
According to the credit risk element decoding method, credit risk element vector data are generated by carrying out credit risk element decoding processing on the bank credit risk scoring data, the credit risk element decoding processing reveals finer granularity information hidden in the credit risk scoring data, so that understanding of credit risks of borrowers is enriched, the credit risk element decoding processing reveals finer granularity information hidden in the credit risk scoring data, understanding of credit risks of borrowers is enriched, credit product characteristic quantum data are generated by utilizing product characteristic quantization processing based on the credit risk element vector data, product characteristic quantization processing can convert multidimensional product characteristic information into concise quantum representation, complexity of the data is reduced, subsequent matching and analysis are facilitated, the quantization processing method can capture implicit association relation among product characteristics, correlation among the product characteristics is accurately reflected in subsequent matching degree calculation, then fusion matching degree among the credit product characteristic quantum data is calculated by using a corresponding fusion matching algorithm (such as similarity calculation or Euclidean distance measurement method), correlation degree among different characteristics can be determined, and importance degree of correlation between the different characteristics is guaranteed, and a basic process is provided for subsequent processing of the credit product characteristics. The credit product matching degree quantum sorting data is generated by carrying out matching degree quantum sorting processing on the credit product fusion matching degree data, the sorting result enables a bank to focus on the credit product with higher matching degree, the complexity of information overload and decision is reduced, the credit product intelligent list generation is carried out on the credit product matching degree quantum sorting data, the bank credit product matching list data is generated, the intelligent list generation is used for customizing an individualized credit product recommendation list for a bank customer according to the demand of borrowers and the matching degree quantum sorting data, and the selection more conforming to the demand and the risk preference is provided.
As an example of the present invention, referring to fig. 4, a detailed implementation step flow diagram of step S3 in fig. 1 is shown, where step S3 includes:
step S31: carrying out credit risk element decoding processing on the credit risk scoring data of the bank to generate credit risk element vector data;
In the embodiment of the invention, a group of random data is selected from a bank rights management data snapshot as seed data, each credit risk rating data is decoded into a credit risk element vector by utilizing a random forest algorithm, a deep learning-based natural language processing model is introduced, the model is trained to understand and identify the credit risk elements in the seed data and decode the credit risk elements into a vector form, credit risk element vector data is generated, the identified credit risk element vector data is input into an automatic encoder, and the high-dimensional representation of the credit risk elements is automatically learned and compressed into a low-dimensional credit risk element vector.
Step S32: generating credit product characteristic quantum data by utilizing product characteristic quantization processing based on credit risk element vector data;
In the embodiment of the invention, credit risk element vector data is converted into a group of initial quantum states by using a quantum state generator. The banking product characteristic data is then encoded into quantum states using quantum state encoding techniques. These banking product characteristics may include, but are not limited to, product interest rate, product deadline, product risk, etc., to generate credit product characteristic quantum data.
Step S33: carrying out fusion matching calculation on credit product characteristic quantum data to generate credit product fusion matching degree data;
In the embodiment of the invention, the credit product characteristic quantum data is calculated by using a corresponding fusion matching algorithm (such as similarity calculation or Euclidean distance measurement method) so as to calculate the matching degree between different credit product characteristics, determine the association degree and importance between the different credit product characteristics, and then fusion the matching degree between the different credit product characteristics by using a weighted summation method to finally generate the credit product fusion matching degree data.
Step S34: carrying out matching degree quantum-level sequencing treatment on the credit product fusion matching degree data to generate credit product matching degree quantum sequencing data;
In the embodiment of the invention, a quantum state entanglement technology is utilized to entanglement a new quantum state with an original quantum state to generate quantum state matching degree data, and a quantum bit exchange algorithm of a quantum computer is utilized to sort the quantum state matching degree data, so that credit product matching measurement sub-sorting data is generated.
Step S35: and carrying out credit product intelligent list matching processing on the credit product matching degree quantum sorting data to generate bank credit product matching list data.
In the embodiment of the invention, the quantum state decoding algorithm on the quantum computer is utilized to decode the credit product matching degree quantum ordering data back to the traditional bit data, and the decoded data is input into a deep reinforcement learning model (such as a model based on a deep Q network) as a state. The model learns the mapping relation from the current state (the matching degree of credit products) to the optimal action (recommended credit products) by continuously interacting with the environment (the information of the credit products), and optimizes the output of the deep reinforcement learning by utilizing collaborative filtering or a content-based recommendation algorithm, so as to generate the matching list data of the bank credit products.
Preferably, the specific steps of step S4 are:
Step S41: performing meta-universe credit landscape modeling processing on the bank credit product matching list data to generate super credit landscape characterization data;
Step S42: heterogeneous role modeling processing is carried out based on the super credit landscape characterization data, and heterogeneous role credit gene data are generated;
step S43: performing heterogeneous character credit behavior simulation processing on heterogeneous character credit gene data to generate heterogeneous character credit behavior simulation field data;
Step S44: carrying out credit behavior pattern pedigree fusion analysis on the heterogeneous character credit behavior simulation field data to generate fusion credit behavior pattern pedigree data;
Step S45: and carrying out credit prediction report generation processing on the fused credit behavior pattern pedigree data to generate bank credit prediction report data.
The invention generates super credit landscape characterization data by performing meta-universe credit landscape modeling processing on the bank credit product matching list data, can creatively characterize the bank credit product matching list data by the meta-universe credit landscape modeling processing, converts the super credit landscape characterization data into the super credit landscape characterization data, can more comprehensively and multidimensional reveal the characteristics and the association of credit data based on the super credit landscape characterization data, performs heterogeneous role modeling processing on the basis of the super credit landscape characterization data, generates heterogeneous role credit gene data, can personally characterize the characteristics and behaviors of different roles in the credit field by the heterogeneous role modeling processing, can accurately predict and evaluate the credit conditions of the different roles, provides support for customized credit decisions, performs heterogeneous role credit behavior simulation processing on the heterogeneous role credit gene data, generates heterogeneous role credit behavior simulation field data by the heterogeneous role credit behavior simulation processing, can generate diversified behavior simulation scene data by simulating credit behaviors of different heterogeneous roles under different situations, including aspects of lending, investment, consumption and the like, thereby providing data guarantee for subsequent analysis and prediction, can form a fused credit pattern by the analysis of the heterogeneous role behavior model, and the analysis of the heterogeneous role credit model by the analysis model, the credit behavior prediction report data is generated by carrying out credit behavior prediction report generation processing on the integrated credit behavior pattern pedigree data, and can not only predict the credit condition of borrowers, but also help banks identify potential risks and provide risk control and optimization suggestions. The information and insight in the report helps the bank to accurately assess risk, formulate risk management strategies, reduce bad asset risk and improve asset quality.
As an example of the present invention, referring to fig. 5, a detailed implementation step flow diagram of step S4 in fig. 1 is shown, where step S4 includes:
Step S41: performing meta-universe credit landscape modeling processing on the bank credit product matching list data to generate super credit landscape characterization data;
In the embodiment of the invention, each bank credit product is regarded as an independent entity, a corresponding virtual credit product is created in the meta universe, each virtual credit product contains all the attributes and characteristics of the corresponding actual credit product, such as product type, credit grade, loan amount, interest rate, repayment term and the like, a virtual space which can contain all the virtual credit products is created, namely the meta universe credit landscape, the positions and interrelationships of the virtual credit products are dynamically adjusted according to the attributes and the characteristics of the virtual credit products, the meta universe credit landscape is molded according to the attributes and the characteristics of the virtual credit products and the interrelationships of the virtual credit products and other virtual credit products, the meta universe credit landscape is encoded, and super credit landscape characterization data is generated.
Step S42: heterogeneous role modeling processing is carried out based on the super credit landscape characterization data, and heterogeneous role credit gene data are generated;
In the embodiment of the invention, heterogeneous roles are defined by various different types of banking clients defined based on factors such as credit history, income level, loan type, repayment capability and the like, a unique code called a credit gene is created for the heterogeneous roles, the credit gene is a complex data structure and can contain various information describing the credit characteristics of the roles, a deep learning model is utilized for modeling each heterogeneous role, super credit landscape characterization data is input into the model, and the credit genes of the roles are output. By training this model, complex relationships between the character's credit genes and the super credit landscape characterization data are learned, and each heterogeneous character will be predicted using the trained model, thereby generating heterogeneous character credit gene data.
Step S43: performing heterogeneous character credit behavior simulation processing on heterogeneous character credit gene data to generate heterogeneous character credit behavior simulation field data;
In the embodiment of the invention, the credit behaviors of different heterogeneous roles are simulated and predicted by using the established heterogeneous role credit gene data, including the credit scores, lending behaviors, consumption habits and the like of the different heterogeneous roles, so as to simulate the credit behavior simulation scenes of the different heterogeneous roles, enable the credit characteristics and behavior patterns of the different heterogeneous roles to be better understood, and finally generate heterogeneous role credit behavior simulation field data.
Step S44: carrying out credit behavior pattern pedigree fusion analysis on the heterogeneous character credit behavior simulation field data to generate fusion credit behavior pattern pedigree data;
In the embodiment of the invention, firstly, the credit behavior simulation field data of the heterogeneous roles are analyzed by using a behavior pattern analysis method to deeply understand the credit behavior patterns of different roles in a simulation scene, including consumption habit, repayment record and the like, the processed data are analyzed and modeled by using a data mining and statistical analysis technology to identify the credit behavior patterns of each heterogeneous role, including behavior characteristics, trends, preferences and the like of each heterogeneous role, meanwhile, the credit behavior patterns of each heterogeneous role are analyzed by using a correlation rule mining and the like to analyze and search the correlation relationship among the credit behavior patterns of different heterogeneous roles, and the credit behavior patterns of each heterogeneous role are evolved by combining the technologies of the correlation relationship of the credit behavior patterns of each heterogeneous role, such as time sequence analysis, machine learning (including long-short-time memory neural network, convolution neural network and the like) to predict the development trend of the credit behavior patterns of each heterogeneous role in the future, and the credit behavior pattern trend of each heterogeneous role is analyzed by using a time-space trajectory simulation analysis method to simulate the evolution behavior patterns of different heterogeneous roles, and the evolution paths of the heterogeneous roles can be known on the evolution paths of the credit behavior patterns of each heterogeneous role. And secondly, analyzing the credit behavior patterns of each heterogeneous character by combining the credit behavior pattern evolution paths of each heterogeneous character to construct an evolution lineage topological structure of the credit behavior patterns of each heterogeneous character, and comprehensively considering the association and development paths among different heterogeneous characters by combining the credit behavior pattern association relations of each heterogeneous character so as to integrate the credit behavior patterns of different heterogeneous characters into a unified lineage, revealing the association, hierarchy and evolution relations among the credit behavior patterns and finally generating the integrated credit behavior pattern lineage data.
Step S45: and carrying out credit prediction report generation processing on the fused credit behavior pattern pedigree data to generate bank credit prediction report data.
In the embodiment of the invention, key characteristics are extracted from the spectrum data of the fused credit behavior mode by utilizing a t distribution neighborhood embedding algorithm, the extracted characteristics comprise the frequency, the intensity, the periodicity and the like of credit behaviors, the extracted characteristics are mapped to a new characteristic space, a credit behavior prediction model is constructed by utilizing a cyclic neural network, the complex relationship between data in the characteristic space and predicted credit behaviors is learned by utilizing a training model, and the data in the mapped characteristic space is predicted by utilizing the trained credit behavior prediction model, so that bank credit behavior prediction report data is generated.
Preferably, the specific steps of step S44 are:
Step S441: carrying out credit behavior pattern analysis on the credit behavior simulation field data of the heterogeneous roles to obtain credit behavior simulation pattern data of each heterogeneous role;
In the embodiment of the invention, firstly, credit behavior simulation field data of heterogeneous roles are analyzed by using a behavior pattern analysis method to deeply understand credit behavior patterns of different roles in a simulation scene, including consumption habits, repayment records and the like, and are subjected to data preprocessing and cleaning to ensure the accuracy and consistency of the data, and then, the processed data are analyzed and modeled by using a data mining and statistical analysis technology to identify the credit behavior patterns of each heterogeneous role, including the behavior characteristics, the trends, the preferences and the like of each heterogeneous role, so that the credit behavior simulation pattern data of each heterogeneous role is finally obtained.
Step S442: carrying out pattern association analysis on the credit behavior simulation pattern data of each heterogeneous character to obtain credit behavior pattern association relation data of each heterogeneous character;
In the embodiment of the invention, the credit behavior simulation mode data of each heterogeneous character is analyzed by using methods such as association rule mining and the like to analyze and explore the association relation among the credit behavior modes of different heterogeneous characters, find out the relativity and influence factors among the credit behavior modes of different heterogeneous characters and finally obtain the credit behavior mode association relation data of each heterogeneous character.
Step S443: carrying out model evolution prediction analysis on the credit behavior simulation model data of each heterogeneous role based on the credit behavior model association relation data of each heterogeneous role to obtain credit behavior model evolution development trend data of each heterogeneous role;
In the embodiment of the invention, the credit behavior simulation mode data of each heterogeneous character is subjected to evolution analysis by combining the credit behavior mode association relation data of each heterogeneous character through technologies such as time sequence analysis, machine learning (including long-short-time memory neural network, convolution neural network and the like) so as to evolve and predict the future credit behavior mode development trend of each heterogeneous character, and the method can help banks to timely adjust credit policies or risk management measures, and finally obtain the credit behavior mode evolution development trend data of each heterogeneous character.
Step S444: performing evolution path simulation analysis on the credit behavior pattern evolution trend data of each heterogeneous role to obtain credit behavior pattern evolution paths of each heterogeneous role;
In the embodiment of the invention, the credit behavior pattern evolution trend data of each heterogeneous character is analyzed by using a time-space evolution track simulation analysis method so as to simulate the credit behavior pattern evolution paths and possible development directions of different heterogeneous characters, know the development tracks of different heterogeneous characters on the credit behavior patterns and finally obtain the credit behavior pattern evolution paths of each heterogeneous character.
Step S445: and carrying out credit behavior pattern pedigree fusion analysis on the credit behavior simulation pattern data of each heterogeneous character based on the credit behavior pattern association relation data of each heterogeneous character and the credit behavior pattern evolution path of each heterogeneous character, and generating fusion credit behavior pattern pedigree data.
In the embodiment of the invention, the credit behavior simulation mode data of each heterogeneous character is firstly analyzed by combining the credit behavior mode evolution path of each heterogeneous character to construct an evolution pedigree topological structure of the credit behavior mode of each heterogeneous character, and then the association and development paths among different heterogeneous characters are comprehensively considered by combining the credit behavior mode association relation data of each heterogeneous character, so that the credit behavior modes of different heterogeneous characters are fused into a unified pedigree, the association, hierarchy and evolution relation among the credit behavior modes are revealed, and finally the fused credit behavior mode pedigree data is generated.
The credit behavior pattern analysis is carried out on the credit behavior simulation field data of the heterogeneous roles, so that the credit behavior patterns of different roles in the simulation scene can be deeply understood, the credit behavior characteristics, trends and preferences of each role can be recognized, and basic data can be provided for subsequent decisions. For example, for the financial arts, such analysis may reveal credit preferences for different customer groups, helping banks or financial institutions to better formulate personalized credit policies. Secondly, by carrying out pattern association analysis on the credit behavior simulation pattern data of each heterogeneous character, the credit behavior pattern association relation data among the heterogeneous characters can be obtained, which is helpful to discover the mutual influence and association among different characters, thereby more accurately predicting the credit behaviors thereof. By carrying out model evolution prediction analysis based on the credit behavior model association relationship data of each heterogeneous role, the future credit behavior model development trend of each role can be presumed, which is helpful for finding potential risks or opportunities as soon as possible, taking corresponding measures, and also helping banks to adjust credit policies or risk management measures in time and reduce bad loan rates. Then, after the evolution path simulation analysis is carried out on the credit behavior pattern evolution trend data of each heterogeneous character, the credit behavior pattern evolution path of each heterogeneous character can be obtained, which is helpful for knowing the development track of different heterogeneous characters on credit behaviors and provides reference for future decisions. Finally, credit behavior pattern pedigree fusion analysis is conducted on the credit behavior simulation pattern data of each heterogeneous character based on the credit behavior pattern association relation data and the credit behavior pattern evolution path of each heterogeneous character, so that the correlation and the development path among different heterogeneous characters are comprehensively considered, the credit behavior patterns of different heterogeneous characters can be integrated into a unified pedigree, and the correlation, the hierarchy and the evolution relation among the different heterogeneous characters are revealed to generate a fusion credit behavior pattern pedigree, so that a more comprehensive view angle and basis can be provided for subsequent credit evaluation, risk management and decision making.
Preferably, the specific steps of step S5 are:
step S51: carrying out credit prediction model construction processing on the credit prediction report data of the bank to generate credit prediction model data;
step S52: optimizing the credit prediction model data by utilizing dynamic credit decision optimization processing to generate credit optimization decision report data;
step S53: and carrying out policy risk assessment policy analysis processing on the credit optimization decision report data to generate bank credit decision report data.
The invention generates credit forecast model data by carrying out credit forecast model construction processing on the credit forecast report data, can more accurately evaluate the credit risk of borrowers by constructing the credit forecast model, identify potential illegal behaviors and bad credit behaviors, help banks to formulate more reasonable credit giving strategies, improve the overall operation efficiency, reduce the labor cost and the time cost, enhance the competitiveness, optimize the credit forecast model data by utilizing dynamic credit decision optimization processing, generate credit optimization decision report data, dynamically adjust and optimize the credit decision according to real-time market change and risk condition, and can dynamically adjust and optimize the credit decision according to the latest credit information and market trend of borrowers, the credit line, the interest rate, the repayment period and the like are adjusted to control risks to the greatest extent, dynamic credit decision optimization can dynamically adjust and optimize credit decisions according to real-time market changes and risk conditions, the credit line, the interest rate, the repayment period and the like can be adjusted according to the latest credit information and market trends of borrowers, the fund utilization efficiency of banks is improved, the risk of credit loss is reduced, risk assessment policy analysis processing is carried out on credit optimization decision report data, bank credit decision report data is generated, and risk assessment policy analysis can provide decision transparency, so that the basis and reason of bank capable of interpreting and interpreting credit decisions can be established for customer trust and satisfaction, and verifiable and interpretable decision processes are provided for a supervision organization.
As an example of the present invention, referring to fig. 6, a detailed implementation step flow diagram of step S5 in fig. 1 is shown, where step S5 includes:
step S51: carrying out credit prediction model construction processing on the credit prediction report data of the bank to generate credit prediction model data;
In the embodiment of the invention, credit decision parameter dynamic adjustment processing is performed on credit prediction model data to generate credit optimization parameter data, credit behavior pattern modeling processing is performed on the credit optimization parameter data to generate credit behavior pattern data, model training optimization processing is performed on the credit behavior pattern data to generate credit optimization model data, model prediction evaluation processing is performed on the credit optimization model data to generate credit prediction model data.
Step S52: optimizing the credit prediction model data by utilizing dynamic credit decision optimization processing to generate credit optimization decision report data;
In the embodiment of the invention, a dynamic credit decision model is constructed by reinforcement learning to realize dynamic optimization of credit decisions, the trained credit behavior prediction model and the dynamic credit decision model are optimized for credit decisions, and the optimization targets include, but are not limited to, minimizing offending risks, maximizing profits and the like, and risk assessment strategy analysis processing is performed on credit optimization decision report data to generate bank credit decision report data.
Step S53: and carrying out risk assessment policy analysis processing on the credit optimization decision report data to generate bank credit decision report data.
In the embodiment of the invention, risk assessment is carried out on the optimized credit decision report data, including potential risks quantification by using risk sensitivity analysis and Monte Carlo simulation, strategy analysis is carried out according to risk assessment results, influences of different credit strategies on bank risk exposure are analyzed, minimum risks of the strategies are adjusted, and a detailed bank credit decision report is generated based on the risk assessment and strategy analysis results.
In one embodiment of the present specification, there is provided a credit authorization data processing system for an integrated credit system, comprising:
at least one processor;
And a memory communicatively coupled to the at least one processor;
Wherein the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the credit authorization data processing method of the integrated credit system as defined in any one of the above.
The invention provides a credit data processing system of a comprehensive credit system, which can realize the credit data processing method of any comprehensive credit system, obtain a bank information center network interface, perform heterogeneous data collection cloud integration processing on the bank information center network interface, generate applicant credit cloud data, perform deep learning risk assessment processing and intelligent credit product matching based on the applicant credit cloud data, generate bank credit product matching list data, perform meta-universe credit simulation processing on the bank credit product matching list data, generate bank credit behavior prediction report data, perform dynamic credit decision optimization processing on the bank credit behavior prediction report data, generate bank credit decision report data, and the system internally follows a set instruction set to complete the operation steps of the method so as to promote completion of the credit data processing method of the comprehensive credit system.
The foregoing is only a specific embodiment of the invention to enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A credit data processing method of a comprehensive credit system is characterized by comprising the following steps:
Step S1: acquiring a bank information center network interface, collecting heterogeneous data of the bank information center network interface, integrating the heterogeneous data in a cloud, and generating credit cloud data of the applicant;
Step S2: performing deep learning risk assessment processing based on the credit cloud data of the applicant to generate bank credit risk scoring data;
Step S3: performing intelligent credit product list matching processing on the bank credit risk scoring data to generate bank credit product matching list data;
Step S4: performing meta-universe credit simulation processing on the matching list data of the bank credit products to generate bank credit behavior prediction report data;
Step S5: and carrying out dynamic credit decision optimization processing on the bank credit behavior prediction report data to generate bank credit decision report data.
2. The method for processing credit authorization data of the integrated credit system according to claim 1, wherein the specific steps of step S1 are as follows:
Step S11: heterogeneous data source acquisition processing is carried out on a network interface of a bank information center, so that an applicant multi-source information set is generated;
step S12: carrying out cloud big data processing based on the applicant multi-source information set to generate applicant cloud information portrait data;
Step S13: carrying out credit evaluation model creation processing on the cloud information portrait data of the applicant to generate credit evaluation model data of the applicant;
Step S14: and carrying out credit cloud data generation processing on the credit evaluation model data of the applicant to generate the credit cloud data of the applicant.
3. The method for processing credit authorization data of the integrated credit system according to claim 1, wherein the specific steps of step S2 are as follows:
step S21: performing data mining and extraction processing on the credit cloud data of the applicant to generate basic information data of the applicant;
step S22: training the basic information data of the applicant by using a deep learning edge training model to generate an applicant trust edge training data set;
Step S23: performing reasoning optimization analysis based on the applicant trust edge training data set to generate applicant trust reasoning data;
Step S24: and carrying out risk assessment processing on the credit reasoning data of the applicant to generate bank credit risk scoring data.
4. The credit data processing method of the integrated credit system according to claim 3, wherein the specific steps of step S23 are as follows:
Step S231: performing trust edge feature analysis on the trust edge training data set of the applicant to obtain trust edge feature data of the applicant;
Step S232: performing trust edge situation feature extraction processing on the trust edge feature data of the applicant to obtain trust edge situation feature information data of the applicant;
Step S233: performing edge situation feature point set sampling processing on the applicant trusted edge situation feature information data to obtain an applicant trusted edge situation feature point set;
step S234: performing edge situation simulation on the applicant trust edge situation feature point set to obtain an applicant trust edge situation simulation space;
Step S235: and carrying out reasoning optimization analysis on the applicant trust edge training data set based on the applicant trust edge situation simulation space to generate applicant trust reasoning data.
5. The method for processing credit data of the integrated credit system according to claim 4, wherein the specific steps of step S235 are:
performing trust edge decision state analysis on the trust edge training data set of the applicant to obtain trust edge decision state data of the applicant;
Performing trust edge decision action analysis on the trust edge training data set of the applicant to obtain trust edge decision action data of the applicant;
Performing credit edge decision rewarding analysis on the credit edge training data set of the applicant to obtain credit edge decision rewarding data of the applicant;
Performing credit-giving edge decision mapping analysis on the credit-giving edge decision state data of the applicant, the credit-giving edge decision action data of the applicant and the credit-giving edge decision rewarding data of the applicant to obtain credit-giving edge decision mapping strategy data of the applicant;
and carrying out reasoning optimization analysis on the applicant trust edge decision mapping strategy data based on the applicant trust edge situation simulation space to generate applicant trust reasoning data.
6. The method for processing credit authorization data of the integrated credit system according to claim 1, wherein the specific steps of step S3 are as follows:
step S31: carrying out credit risk element decoding processing on the credit risk scoring data of the bank to generate credit risk element vector data;
step S32: generating credit product characteristic quantum data by utilizing product characteristic quantization processing based on credit risk element vector data;
Step S33: carrying out fusion matching calculation on credit product characteristic quantum data to generate credit product fusion matching degree data;
Step S34: carrying out matching degree quantum-level sequencing treatment on the credit product fusion matching degree data to generate credit product matching degree quantum sequencing data;
step S35: and carrying out credit product intelligent list matching processing on the credit product matching degree quantum sorting data to generate bank credit product matching list data.
7. The method for processing credit authorization data of the integrated credit system according to claim 1, wherein the specific steps of step S4 are as follows:
Step S41: performing meta-universe credit landscape modeling processing on the bank credit product matching list data to generate super credit landscape characterization data;
Step S42: heterogeneous role modeling processing is carried out based on the super credit landscape characterization data, and heterogeneous role credit gene data are generated;
step S43: performing heterogeneous character credit behavior simulation processing on heterogeneous character credit gene data to generate heterogeneous character credit behavior simulation field data;
Step S44: carrying out credit behavior pattern pedigree fusion analysis on the heterogeneous character credit behavior simulation field data to generate fusion credit behavior pattern pedigree data;
Step S45: and carrying out credit prediction report generation processing on the fused credit behavior pattern pedigree data to generate bank credit prediction report data.
8. The method for processing credit authorization data of the integrated credit system according to claim 7, wherein the specific steps of step S44 are as follows:
Step S441: carrying out credit behavior pattern analysis on the credit behavior simulation field data of the heterogeneous roles to obtain credit behavior simulation pattern data of each heterogeneous role;
step S442: carrying out pattern association analysis on the credit behavior simulation pattern data of each heterogeneous character to obtain credit behavior pattern association relation data of each heterogeneous character;
Step S443: carrying out model evolution prediction analysis on the credit behavior simulation model data of each heterogeneous role based on the credit behavior model association relation data of each heterogeneous role to obtain credit behavior model evolution development trend data of each heterogeneous role;
step S444: performing evolution path simulation analysis on the credit behavior pattern evolution trend data of each heterogeneous role to obtain credit behavior pattern evolution paths of each heterogeneous role;
Step S445: and carrying out credit behavior pattern pedigree fusion analysis on the credit behavior simulation pattern data of each heterogeneous character based on the credit behavior pattern association relation data of each heterogeneous character and the credit behavior pattern evolution path of each heterogeneous character, and generating fusion credit behavior pattern pedigree data.
9. The method for processing credit authorization data of the integrated credit system according to claim 1, wherein the specific steps of step S5 are as follows:
step S51: carrying out credit prediction model construction processing on the credit prediction report data of the bank to generate credit prediction model data;
step S52: optimizing the credit prediction model data by utilizing dynamic credit decision optimization processing to generate credit optimization decision report data;
step S53: and carrying out policy risk assessment policy analysis processing on the credit optimization decision report data to generate bank credit decision report data.
10. A credit data processing system for an integrated credit system, comprising:
at least one processor;
a memory communicatively coupled to the at least one processor;
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the credit data processing method of the integrated credit system of any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410423023.7A CN118037440A (en) | 2024-04-09 | 2024-04-09 | Trusted data processing method and system for comprehensive credit system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410423023.7A CN118037440A (en) | 2024-04-09 | 2024-04-09 | Trusted data processing method and system for comprehensive credit system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118037440A true CN118037440A (en) | 2024-05-14 |
Family
ID=90989530
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410423023.7A Pending CN118037440A (en) | 2024-04-09 | 2024-04-09 | Trusted data processing method and system for comprehensive credit system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118037440A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130091050A1 (en) * | 2011-10-10 | 2013-04-11 | Douglas Merrill | System and method for providing credit to underserved borrowers |
| CN112991041A (en) * | 2021-02-23 | 2021-06-18 | 深圳季连科技有限公司 | Credit line sharing method, device, terminal and storage medium based on big data |
| CN114661455A (en) * | 2020-12-22 | 2022-06-24 | 英特尔公司 | Method and apparatus for validating trained models in edge environments |
| CN116188140A (en) * | 2023-01-04 | 2023-05-30 | 武汉众邦银行股份有限公司 | Credit financial risk policy production method and device |
| CN116579852A (en) * | 2023-06-29 | 2023-08-11 | 中国工商银行股份有限公司 | Financial service providing method, device and storage medium based on meta universe |
-
2024
- 2024-04-09 CN CN202410423023.7A patent/CN118037440A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130091050A1 (en) * | 2011-10-10 | 2013-04-11 | Douglas Merrill | System and method for providing credit to underserved borrowers |
| CN114661455A (en) * | 2020-12-22 | 2022-06-24 | 英特尔公司 | Method and apparatus for validating trained models in edge environments |
| CN112991041A (en) * | 2021-02-23 | 2021-06-18 | 深圳季连科技有限公司 | Credit line sharing method, device, terminal and storage medium based on big data |
| CN116188140A (en) * | 2023-01-04 | 2023-05-30 | 武汉众邦银行股份有限公司 | Credit financial risk policy production method and device |
| CN116579852A (en) * | 2023-06-29 | 2023-08-11 | 中国工商银行股份有限公司 | Financial service providing method, device and storage medium based on meta universe |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liang et al. | Adversarial deep reinforcement learning in portfolio management | |
| Zhou et al. | Stock market prediction on high‐frequency data using generative adversarial nets | |
| CN112819604A (en) | Personal credit evaluation method and system based on fusion neural network feature mining | |
| KR102330423B1 (en) | Online default forecasting system using image recognition deep learning algorithm | |
| CN114219360A (en) | Monitoring safety prediction method and system based on model optimization | |
| CN117349782B (en) | Intelligent data early warning decision tree analysis method and system | |
| CN113011895A (en) | Associated account sample screening method, device and equipment and computer storage medium | |
| CN116861331A (en) | Expert model decision-fused data identification method and system | |
| CN117391247A (en) | Enterprise risk level prediction method and system based on deep learning | |
| Sujatha et al. | Loan prediction using machine learning and its deployement on web application | |
| Wu et al. | An uncertainty-oriented cost-sensitive credit scoring framework with multi-objective feature selection | |
| CN117808214A (en) | Hydraulic engineering data analysis system | |
| US20210090101A1 (en) | Systems and methods for business analytics model scoring and selection | |
| CN116911994B (en) | External trade risk early warning system | |
| CN114693307B (en) | Securities futures programmed trading strategy risk pressure test system | |
| CN118037440A (en) | Trusted data processing method and system for comprehensive credit system | |
| Zang | Construction of Mobile Internet Financial Risk Cautioning Framework Based on BP Neural Network | |
| CA3160715A1 (en) | Systems and methods for business analytics model scoring and selection | |
| KAMLEY et al. | Multiple regression: A data mining approach for predicting the stock market trends based on open, close and high price of the month | |
| CN117557361B (en) | User credit risk assessment method and system based on data analysis | |
| İNAL | Use of artificial intelligence in fintech tools in terms of risk management | |
| Bayuk et al. | The research on stability of the Russian banking system by machine learning methods | |
| CN116258574B (en) | Mixed effect logistic regression-based default rate prediction method and system | |
| CN117649236A (en) | Risk prediction method, apparatus and storage medium for transaction | |
| Acheampong et al. | Supporting Environmental Decision MakingApplication of Machine Learning Techniques to Australia’s Emissions |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |