Iterative Factor Analysis for Valid Risk Evaluation Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional unsupervised learning techniques for risk evaluation struggle to eliminate inappropriate risk evaluation values, leading to inaccurate comprehensive risk values due to strong dependence on specific viewpoints or situations.
Innovation Solution
An evaluation program and method that performs factor analysis to calculate first factor loadings for multiple risk evaluation values, excluding those with loadings less than a predetermined value, and iteratively refining the analysis until no inappropriate values remain, thereby improving the validity of the comprehensive risk evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If unsupervised learning is used to calculate comprehensive risk evaluation values from multiple techniques, then the calculation can be performed automatically without manual intervention, but inappropriate risk evaluation values cannot be eliminated leading to reduced accuracy
Solution Approach 1:
The patent implements an iterative feedback mechanism where factor analysis results are used to identify and remove inappropriate risk evaluation values, then the analysis is repeated on the refined dataset. This feedback loop continues until convergence, automatically eliminating inaccurate values while maintaining computational automation.
Solution Approach 2:
The patent performs preliminary factor analysis before final comprehensive risk evaluation to identify and remove inappropriate risk evaluation values in advance. This preliminary action ensures that only suitable values are incorporated into the final calculation, improving accuracy while maintaining automation.
2Adaptability or versatility
If all risk evaluation values from multiple techniques are incorporated into the comprehensive evaluation, then the evaluation covers all viewpoints, but inappropriate values with strong dependence on specific viewpoints degrade the overall validity
Solution Approach 1:
The patent extracts and removes inappropriate risk evaluation values from the dataset based on factor loading thresholds. By taking out these problematic values that show strong dependence on specific viewpoints, the remaining comprehensive evaluation maintains both versatility and reliability.
Solution Approach 2:
The patent changes the parameter set by dynamically adjusting which risk evaluation values are included based on their factor loadings. Values with loadings below the threshold are excluded, transforming the parameter set to include only reliable viewpoints while maintaining comprehensiveness.
3Reliability
If factor analysis is performed iteratively to remove inappropriate values, then the validity of risk evaluation is improved, but the calculation process becomes more complex
Solution Approach 1:
The iterative factor analysis process is self-service in nature, automatically identifying and removing inappropriate values without requiring external intervention or complex manual procedures. The system serves itself by using its own output to refine its input, simplifying the overall process despite the iterative nature.
Solution Approach 2:
The feedback mechanism guides the iterative process to converge automatically, preventing infinite loops and reducing the need for complex control logic. Each iteration refines the dataset based on previous results, with the process self-terminating when convergence criteria are met.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An evaluation program causes a computer to calculate a first factor loading of each of a plurality of risk evaluation values by performing factor analysis on the plurality of risk evaluation values each calculated according to one of a plurality of techniques. The evaluation program then causes the computer to exclude, from the plurality of risk evaluation values, a risk evaluation value for which the calculated first factor loading is less than a predetermined value. Then, the evaluation program causes the computer to perform the factor analysis on the plurality of risk evaluation values obtained by excluding the risk evaluation value for which the first factor loading is equal to or less than the predetermined value. Then, the evaluation program causes the computer to repeat the processing described above until no risk evaluation value for which the first factor loading is equal to or less than the predetermined value appears by the factor analysis.