Compound Model for Prediction Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data mining prediction methods in finance and insurance rely on single models, which have limited explanation angles and high prediction error rates due to their inherent limitations.
Innovation Solution
A method and device that train multiple pre-determined models (such as decision tree, linear regression, and neural networks) on client information samples and combine them into a compound model using specific combination rules to enhance prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single model is used for prediction, then the device complexity is low, but the prediction accuracy is limited and error rate is high
Solution Approach 1:
The patent combines multiple different prediction models (e.g., logistic regression model, decision tree model, neural network model) into a single compound model. Each model has its own prediction result, and these results are integrated through a combination mechanism (such as weighted averaging or voting) to produce the final prediction outcome. This merging approach leverages the complementary strengths of different models to improve overall prediction accuracy while managing complexity through systematic integration.
2Measurement precision
If multiple models are combined into a compound model, then the prediction accuracy improves, but the device complexity increases
Solution Approach 1:
The compound model is designed to perform multiple functions by integrating different types of prediction models, each contributing unique capabilities. The system can handle various prediction tasks using the same compound model structure, making it universally applicable. The combination rule mechanism allows the system to adaptively weight or select different model outputs based on the specific prediction scenario, achieving multi-functionality without proportionally increasing complexity.
Solution Approach 2:
The patent employs parameter adjustment mechanisms within the compound model, such as dynamically adjusting the weights of individual model predictions or modifying combination rules based on performance metrics. By changing parameters like model weights, threshold values, or combination strategies, the system can optimize prediction accuracy for different scenarios without fundamentally altering the compound model structure, thus managing complexity through parameter tuning rather than structural complexity.
Data Source
AI summary
The disclosure discloses a method and device of analysis based on a model, and a computer readable storage medium. The method includes: training various pre-determined models based on a preset number of customer information samples; combining the various trained models into a compound model according to a pre-determined combining rule, and after customer information to be analyzed is received, inputting the customer information to be analyzed into the compound model to output an analysis result. According to the disclosure, by the use of the compound model combined by the various models for analysis and prediction, the advantages of different models can be combined. Compared with a single model for prediction, the compound model effectively improves the accuracy of a prediction result.