Model Setting Support Device for Product Performance Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing model setting methods lack validation of selected models in consideration of product knowledge and are limited by the use of predetermined data analysis methods, which do not ensure appropriate model setting for predicting product performance.
Innovation Solution
A model setting support device and method that includes multiple learning models for estimating product performance, with analysis units to calculate prediction accuracy and contribution of explanatory variables, allowing for the selection of appropriate models and variables based on user input and knowledge of product conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple machine learning algorithms are executed independently to select a model with highest prediction performance, then prediction accuracy is improved, but the process lacks validation considering product knowledge and does not ensure appropriate model setting
Solution Approach 1:
The patent implements feedback by having the analysis unit evaluate learned models using two analysis methods: one calculating prediction accuracy and another assessing model validity based on product knowledge. The learning result output unit provides feedback information about both accuracy and validity, allowing the system to iteratively improve model selection by considering both quantitative performance and qualitative appropriateness.
Solution Approach 2:
The analysis unit acts as an intermediary between the model learning process and the final model selection. It introduces an additional evaluation layer that assesses model validity using product knowledge, bridging the gap between pure prediction performance and appropriate model setting for the specific application domain.
2Ease of manufacture
If predetermined data analysis methods are used for dataset analysis, then analysis process is simplified, but the validity of the data analysis method is not determined and appropriate model setting cannot be ensured
Solution Approach 1:
The patent makes the analysis process dynamic by allowing the selection and application of multiple different analysis methods depending on the specific learning model and product domain. Rather than using a single predetermined method, the system can adaptively choose appropriate analysis approaches, balancing simplicity with validity assessment.
3Measurement precision
If multiple learning models are learned and evaluated, then model selection accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent applies partial action by evaluating multiple models but using two analysis methods to efficiently filter and rank them. The first analysis method quickly assesses prediction accuracy, while the second method provides validity assessment. This layered approach allows the system to process multiple models without fully exhaustive evaluation of each, reducing overall processing time while maintaining selection accuracy.
Data Source
AI summary
A model setting support device includes a model learning unit having a plurality of learning models for estimating an objective variable indicating performance of a product from an explanatory variable group and configured to learn each of the plurality of learning models using training data including a dataset including the explanatory variable group and a target value of the objective variable, an analysis unit configured to execute first analysis for calculating a predicted value of the objective variable from the explanatory variable group and calculating prediction accuracy of the predicted value on the basis of the target value of the objective variable and second analysis for deriving a degree of contribution of the explanatory variable to the objective variable, for each learning model, and a learning result processing unit configured to output learning result information indicating the prediction accuracy and the contribution degree for each learning model.