Sparse Modeling Variable Selection Using Combined Importance Scores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Different sparse modeling methods select varying explanatory variables from the same input data, potentially missing important variables, making it challenging to reliably choose suitable explanatory variables for machine learning models in industrial applications.
Innovation Solution
A machine learning device and method that acquire individual importance degrees using multiple sparse modeling methods and calculate a comprehensive importance degree to select the most suitable explanatory variables, ensuring consistency and accuracy in variable selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single sparse modeling method is used to select explanatory variables, then the selection process is simple and fast, but the reliability of variable selection deteriorates because important variables may be missed
Solution Approach 1:
The patent combines multiple sparse modeling methods (lasso, stepwise, and others) into a unified variable selection system. Each method generates importance degrees for explanatory variables, and these are integrated through a selection unit that aggregates the results. This merging approach ensures that variables selected by any of the individual methods are considered, thereby improving reliability while maintaining computational efficiency through parallel processing.
Solution Approach 2:
The selection unit serves as a universal component that processes importance degrees from multiple different sparse modeling methods. It implements a multi-functional approach by accommodating various modeling techniques and synthesizing their outputs into a comprehensive variable selection result, making the system adaptable to different modeling scenarios while ensuring consistent and reliable variable selection.
2Reliability
If multiple sparse modeling methods are used to improve variable selection reliability, then the accuracy of selecting important variables improves, but the system complexity increases
Solution Approach 1:
The selection unit acts as an intermediary component that mediates between multiple sparse modeling methods and the final variable selection output. It receives importance degrees from various methods, standardizes and aggregates them, and produces a unified selection result. This intermediary structure simplifies the overall system architecture by providing a single point of integration, thereby reducing the perceived complexity despite using multiple modeling methods.
Solution Approach 2:
The system changes parameters by adjusting the number and type of sparse modeling methods employed. The selection unit can flexibly incorporate different methods based on specific application requirements, allowing the system complexity to be tuned. By parameterizing the selection process, the system can maintain reliability through multiple methods while managing complexity through configurable design.
Data Source
AI summary
A machine learning device includes a sparse modeling processing unit and a selection unit. The sparse modeling processing unit acquires individual importance degrees for each of explanatory variable candidates, the individual importance degrees being acquired by using respective sparse modeling methods different from each other, each of the sparse modeling methods taking input data including a specified objective variable in a learning model used for industrial activity and the explanatory variable candidates that are candidates for an explanatory variable for explaining the specified objective variable. The selection unit calculates a comprehensive importance degree for each of the explanatory variable candidates based on the individual importance degrees of each of the explanatory variable candidates, and selects an explanatory variable of the learning model from among the explanatory variable candidates based on the comprehensive importance degree.