Loan approval system and method based on gradient neural tree algorithm
By adopting the gradient neural tree algorithm in the loan approval system, the existing system's inefficiency and bias problems in large-scale and high-dimensional data processing are solved, and efficient and fair loan approval decisions are achieved.
Patent Information
- Application Number
- CN202411289329.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-14
- Publication Date
- 2024-12-06
AI Technical Summary
The existing loan approval system has low computational efficiency, insufficient generalization capabilities of model when processing large-scale and high-dimensional data, lacks transparency and interpretability, and has algorithmic biases and risk assessment bias, resulting in inefficient and unfair decision-making.
The loan approval system based on the gradient neural tree algorithm is adopted, including the user interface module, the gradient neural tree algorithm model module, the business logic module and the data layer module. Through the gradient enhancement tree training model, business rules are integrated, and data preprocessing and model deployment are carried out. Provides automated approval processes and efficient data processing to ensure transparency and fairness of the model.
It significantly improves approval efficiency and accuracy, reduces artificial bias, improves the transparency and interpretability of the model, reduces risk assessment bias, and achieves more objective and fair loan approval decisions.
Smart Images

Figure CN119090627A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of financial technology, and specifically relates to a loan approval system and method based on a gradient neural tree algorithm. Background Art
[0002] With the rapid development of the financial industry, the demand for loan approval is increasing. The current loan approval process mostly relies on manual review and simple scoring models, which is time-consuming, labor-intensive, inefficient and error-prone, and cannot cope with large-scale loan application processing.
[0003] With the development of big data and machine learning technology, model-based automatic approval systems have gradually become mainstream. However, existing systems still have problems such as low computational efficiency and insufficient model generalization ability when processing large-scale, high-dimensional data.
[0004] Specifically, the existing solutions still have the following problems:
[0005] Lack of transparency and explainability: Complex machine learning models can be difficult to explain, making it difficult for customers and regulators to understand the logic behind lending decisions.
[0006] Algorithmic bias: If the data is biased, the algorithm may also be biased, which may lead to unfair treatment of certain groups.
[0007] Risk assessment bias: Loan approval algorithms may be too conservative or aggressive, resulting in good customers being rejected or bad loans being approved. Sometimes algorithms may over-rely on certain indicators (such as credit scores) and ignore other important factors. Summary of the invention
[0008] In view of the above problems, the present invention is proposed to provide a loan approval system and method based on a gradient neural tree algorithm that overcomes the above problems or at least partially solves the above problems.
[0009] To achieve the above object, the present invention adopts the following technical solutions:
[0010] A loan approval system based on a gradient neural tree algorithm, the system comprising:
[0011] A user interface module, used for providing a user interface;
[0012] Gradient neural tree algorithm model module, used to train loan approval model;
[0013] Business logic module, used to integrate business rules;
[0014] The data layer module is used to collect and pre-process data;
[0015] The service layer module is used to receive requests and return approval results.
[0016] Optionally, the user interface module includes:
[0017] The front-end application unit is used to provide a loan application interface, including filling out application forms, uploading documents, and displaying approval results;
[0018] The backend management unit is used to monitor system status, manage models, update the system, and view approval results.
[0019] Optionally, the gradient neural tree algorithm model module includes:
[0020] Model training unit, used to train loan approval model;
[0021] Model evaluation unit, used to evaluate the performance of loan approval models;
[0022] Hyperparameter tuning unit, used to adjust the parameters of the loan approval model;
[0023] Model validation unit, used to evaluate the generalization ability of the loan approval model;
[0024] The model deployment unit is used to deploy the trained loan approval model to the production environment.
[0025] Alternatively, data collection channels include internal bank databases, credit rating agencies and public records.
[0026] Optionally, data preprocessing includes cleaning data, handling missing values, and performing feature engineering.
[0027] Optionally, the service layer modules include:
[0028] Encapsulate the loan approval model as a RestFul API or gRPC service, receive requests from the front end, and return the approval result;
[0029] Use a load balancer to distribute requests;
[0030] Introduce fault-tolerant mechanisms and design fault recovery and redundancy strategies.
[0031] The present invention also provides a loan approval method based on a gradient neural tree algorithm, the method is based on any of the above-mentioned systems, and the method comprises the following steps:
[0032] S1. Fill in personal information and information;
[0033] S2. Based on the filled information, risk assessment and credit scoring are performed based on the gradient neural tree algorithm;
[0034] S3. Recommend products and services to users based on the evaluation results.
[0035] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:
[0036] 1. The present invention greatly improves the approval efficiency: through the automated approval process, the time and workload of manual review can be significantly reduced, rapid response can be achieved, and customer experience can be improved; using the gradient neural tree model, the credit risk of the borrower can be more accurately assessed, the interference of human factors can be reduced, and the accuracy of approval can be improved; by analyzing a large amount of historical data, the gradient neural tree algorithm can identify more risk factors, help banks better manage risks, reduce bad debt rates, and greatly reduce credit risks.
[0037] 2. The present invention adopts the gradient neural tree algorithm to train the model, and utilizes its efficient data processing capability and automatic feature selection, which not only minimizes the model training time, but also greatly improves the prediction accuracy; and the machine learning model can continuously learn and adapt to new data. As the market and customer behavior change, the model can be adjusted and updated to maintain the timeliness of the approval strategy.
[0038] 3. The present invention improves the transparency and interpretability of the model by introducing the gradient neural tree algorithm, ensures the balance and fairness of the algorithm, avoids algorithm bias, and reduces the deviation of risk assessment. The algorithm is based on data-driven decision-making, reduces human bias and subjective judgment, and makes loan approval more objective and fair. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 A schematic diagram of a loan approval system based on a gradient neural tree algorithm provided in an embodiment of the present application;
[0040] Figure 2 A flowchart of a loan approval method based on a gradient neural tree algorithm is provided in an embodiment of the present application. DETAILED DESCRIPTION
[0041] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0042] Embodiment 1:
[0043] See also Figure 1 This embodiment provides a loan approval system based on a gradient neural tree algorithm, the system comprising:
[0044] The user interface module is used to provide a user interface.
[0045] The user interface module includes:
[0046] The front-end application unit is used to provide customers with a loan application interface, including functions such as filling out application forms, uploading documents, and displaying approval results.
[0047] The backend management unit provides an interface for managers to monitor system status, manage models, update the system, and view approval results.
[0048] Gradient neural tree algorithm model module, used to train loan approval model.
[0049] The gradient neural tree algorithm model module includes:
[0050] The model training unit is used to train the loan approval model. The gradient neural tree algorithm is used to train the loan approval model. The model training can use historical loan data, including the basic information, financial status, credit history, etc. of the loan applicant as features, and whether the loan is in default as a label.
[0051] The model evaluation unit is used to evaluate the performance of the loan approval model to ensure its accuracy and reliability.
[0052] Hyperparameter tuning unit, used to adjust the parameters of the loan approval model; model parameters are adjusted through network search, random search and Bayesian optimization to optimize model performance.
[0053] The model validation unit is used to evaluate the generalization ability of the loan approval model. It uses methods such as cross-validation to evaluate the generalization ability of the model and ensure the performance of the model on unknown data.
[0054] The model deployment unit is used to deploy the trained loan approval model to the production environment for real-time evaluation.
[0055] The business logic module is used to integrate business rules, and the approval rule engine is used to integrate business rules, such as credit score thresholds, loan amount calculations, etc. The decision flow is used to make the final loan approval decision based on the model evaluation results and business rules.
[0056] The data layer module is used to collect and pre-process data;
[0057] Channels for collecting data include internal bank databases, credit rating agencies and public records.
[0058] Data preprocessing includes cleaning data, handling missing values, and performing feature engineering. The processed data is stored in a data warehouse or data lake to facilitate model training and real-time query.
[0059] The service layer module is used to receive requests and return approval results.
[0060] The service layer modules include:
[0061] The API service encapsulates the loan approval model as a RestFul API or gRPC service, receives requests from the front end, and returns the approval results.
[0062] Use a load balancer to distribute requests to ensure system stability and responsiveness.
[0063] Introduce fault-tolerant mechanisms and design fault recovery and redundancy strategies to ensure high availability of services.
[0064] Gradient Boosting Tree is a machine learning algorithm based on Gradient Boosting Decision Tree. It is an efficient and fast gradient boosting framework developed by Microsoft. It is specifically used to solve problems such as classification and regression. It was developed by the Microsoft team to provide an efficient, fast and accurate implementation of the gradient boosting algorithm. Compared with traditional gradient boosting methods, LightGBM has advantages in many aspects, including efficiency, speed, accuracy and support for parallelization.
[0065] This embodiment greatly improves the approval efficiency: through the automated approval process, the time and workload of manual review can be significantly reduced, rapid response can be achieved, and customer experience can be improved; using the gradient neural tree model, the credit risk of the borrower can be more accurately assessed, the interference of human factors can be reduced, and the accuracy of approval can be improved; by analyzing a large amount of historical data, the gradient neural tree algorithm can identify more risk factors, help banks better manage risks, reduce bad debt rates, and greatly reduce credit risks.
[0066] This embodiment adopts the gradient neural tree algorithm to train the model, and utilizes its efficient data processing capabilities and automatic feature selection to not only minimize the model training time, but also greatly improve the prediction accuracy; and the machine learning model can continuously learn and adapt to new data. As the market and customer behavior change, the model can be adjusted and updated to maintain the timeliness of the approval strategy.
[0067] This embodiment improves the transparency and interpretability of the model by introducing the gradient neural tree algorithm, ensures the balance and fairness of the algorithm, avoids algorithm bias, and reduces the deviation of risk assessment. The algorithm is based on data-driven decision-making, reduces human bias and subjective judgment, and makes loan approval more objective and fair.
[0068] Embodiment 2:
[0069] like Figure 2As shown, Embodiment 2 discloses a loan approval method based on a gradient neural tree algorithm. The method is based on the above-mentioned loan approval system based on a gradient neural tree algorithm. The method includes the following steps:
[0070] S1. Fill in personal information and information.
[0071] S2. Based on the filled-in information, risk assessment and credit scoring are performed based on the gradient neural tree algorithm.
[0072] S3. Recommend products and services to users based on the evaluation results.
[0073] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.
Claims
1. A loan approval system based on a gradient neural tree algorithm, characterized in that: The system comprises: A user interface module, used for providing a user interface; Gradient neural tree algorithm model module, used to train loan approval model; Business logic module, used to integrate business rules; The data layer module is used to collect and pre-process data; The service layer module is used to receive requests and return approval results.
2. A loan approval system based on a gradient neural tree algorithm as claimed in claim 1, characterized in that: The user interface module includes: The front-end application unit is used to provide a loan application interface, including filling out application forms, uploading documents, and displaying approval results; The backend management unit is used to monitor system status, manage models, update the system, and view approval results.
3. A loan approval system based on a gradient neural tree algorithm as claimed in claim 1, characterized in that: The gradient neural tree algorithm model module includes: Model training unit, used to train loan approval model; Model evaluation unit, used to evaluate the performance of loan approval models; Hyperparameter tuning unit, used to adjust the parameters of the loan approval model; Model validation unit, used to evaluate the generalization ability of the loan approval model; The model deployment unit is used to deploy the trained loan approval model to the production environment.
4. A loan approval system based on a gradient neural tree algorithm as claimed in claim 1, characterized in that: Data collection channels include internal bank databases, credit rating agencies and public records.
5. A loan approval system based on a gradient neural tree algorithm as claimed in claim 1, characterized in that: Data preprocessing includes cleaning data, handling missing values, and performing feature engineering.
6. A loan approval system based on a gradient neural tree algorithm as claimed in claim 1, characterized in that: The service layer modules include: Encapsulate the loan approval model as a RestFul API or gRPC service, receive requests from the front end, and return the approval result; Use a load balancer to distribute requests; Introduce fault-tolerant mechanisms and design fault recovery and redundancy strategies.
7. A loan approval method based on a gradient neural tree algorithm, the method is based on a loan approval system based on a gradient neural tree algorithm as claimed in any one of claims 1 to 6, characterized in that: The method comprises the following steps: S1. Fill in personal information and information; S2. Based on the filled information, risk assessment and credit scoring are performed based on the gradient neural tree algorithm; S3. Recommend products and services to users based on the evaluation results.
