Factor Analysis Grouping for Hidden Influence Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing factor analysis methods struggle to accurately identify hidden effect factors due to the dominance of strong effect factors, leading to underestimated influence degrees for items with strong associations and difficulties in calculating associations without temporal order or clear causal relationships.
Innovation Solution
A factor analysis device and method that classify data into first and second groups, calculating influence degrees specifically for the second group on the first group, allowing for the identification of effect factors that may be masked by strong effect factors and enabling more precise analysis of uncontrollable items.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all items are used as explanatory variables in factor analysis, then the analysis covers all potential effect factors, but the contribution rate of hidden effect factors becomes small and they cannot be accurately specified
Solution Approach 1:
The patent segments the set of all explanatory variables into two groups: (1) variables with remarkably strong influence on the response variable, and (2) remaining variables. By performing factor analysis separately on the second group using variables from the first group as controls, the patent isolates the contribution of hidden effect factors that would otherwise be masked by dominant variables. This segmentation enables accurate specification of effect factors with small contribution rates.
2Measurement precision
If only items with no remarkably strong influence are used for analysis, then hidden effect factors can be identified, but the analysis function does not sufficiently fit and precise analysis cannot be performed
Solution Approach 1:
The patent uses variables with remarkably strong influence as intermediary control variables when analyzing the relationship between hidden effect factors and the response variable. By incorporating these strong-influence variables into the analysis model as controls, the patent ensures that the analysis function maintains sufficient fit and reliability while simultaneously enabling the detection of hidden effect factors that would otherwise be undetectable.
3Measurement precision
If multiple regression analysis is performed to find hidden effect factors, then some hidden factors can be identified, but explanatory variables with large association degree with strong effect factors are still difficult to find
Solution Approach 1:
The patent extracts and removes variables with remarkably strong influence from the set of explanatory variables before performing factor analysis. By taking out these dominant variables that create large association degrees, the patent eliminates the masking effect they have on hidden effect factors. This extraction enables the detection of hidden effect factors that have large association degrees with strong effect factors, which would be difficult to identify in a combined analysis.
Data Source
AI summary
Provided is a factor analysis device capable of obtaining more useful knowledge relating to the degree of influence of pieces of data. A factor analysis device according to one embodiment of the present invention is provided with: a classification unit for classifying a type of data into a first group or a second group; and an influence degree calculation unit for calculating, as the degree of influence on target data, the degree of influence of the data of the type classified into the second group on the data of the first group type.