Regression Analysis for Variable Selection in EV Motion Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately identify and separate influential physical variables affecting the motion of moving objects, such as electric vehicles, due to difficulties in dividing explanatory variable spaces into regions based on variable influence and removing unnecessary variables.
Innovation Solution
A computer-readable recording medium and analysis apparatus that generates a regression function using observed physical variables, divides the explanatory variable space into regions, and provides contribution information on each variable's influence, enabling the identification and selection of key variables for each region.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If regression analysis is performed using all observed physical variables, then comprehensive analysis is achieved, but computational complexity and difficulty in identifying key variables increases
Solution Approach 1:
The explanatory variable space is divided into multiple regions based on the regression function, allowing variables to be analyzed and selected separately for each region. This segmentation enables identification of key variables specific to each operational condition without being overwhelmed by the entire variable set simultaneously.
Solution Approach 2:
The patent extracts and identifies only the necessary explanatory variables that significantly contribute to the objective variable within each region. By removing unnecessary variables through contribution analysis, the system achieves accurate estimation while reducing analysis complexity.
2Measurement precision
If explanatory variable space is divided into multiple regions, then variable influence can be accurately identified, but processing time and computational resources increase
Solution Approach 1:
The regression function is generated in advance using all observed physical variables before region division. This preliminary action creates a foundation that guides the subsequent region division and variable selection processes, making them more efficient and targeted rather than requiring repeated comprehensive analyses.
Solution Approach 2:
Different sets of explanatory variables are identified for different regions of the explanatory variable space. Each region receives tailored variable selection based on its specific characteristics, improving accuracy for local conditions without requiring uniform complex analysis across all possible variable combinations.
3Adaptability or versatility
If all explanatory variables are used for regression analysis, then model completeness is maintained, but model accuracy for specific conditions decreases
Solution Approach 1:
The patent segments the explanatory variable space into multiple regions and selects appropriate explanatory variables for each region. This allows the model to be complete across all conditions while being optimized for accuracy within each specific condition, resolving the conflict between universality and precision.
Solution Approach 2:
The set of explanatory variables is made dynamic rather than static - different variables are selected depending on which region of the explanatory variable space the current data point falls into. This dynamic adaptation maintains model completeness while improving accuracy for specific conditions.
Data Source
AI summary
A non-transitory computer-readable recording medium having stored therein an analysis program causing a computer to execute processing includes: generating a regression function including a plurality of explanatory variables and an objective variable based on regression analysis using observed values of a plurality of physical elements each related to motion of a moving object; dividing an explanatory variable space containing the plurality of explanatory variables into a plurality of regions by using the regression function; and generating contribution information on a contribution of each of the plurality of explanatory variables to the objective variable for each of the plurality of regions.