Factor Variable Grid Analysis for Interpretable Quality Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning models for factor analysis struggle to interpret how to control factor variables to reduce defects, making it difficult to set effective ranges for improving product quality.
Innovation Solution
A method and system for setting factor variable areas by dividing the factor variable space into grids, calculating good densities, and selecting candidate areas based on these densities to define ranges for factor variables, which can be used to improve product quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a machine learning model is used for factor analysis, then analysis accuracy is improved, but interpretability deteriorates making it difficult to control factor variables
Solution Approach 1:
The patent segments the continuous factor variable space into discrete grids, transforming the complex continuous optimization problem into a manageable discrete classification problem. Each grid represents a specific range of factor variable values, and the system calculates good density for each grid to identify optimal regions. This segmentation enables both high accuracy in identifying good quality regions and interpretability in understanding which factor variable ranges lead to good outcomes.
2Productivity
If the range of factor variables is not properly controlled, then productivity is maintained, but manufacturing precision deteriorates
Solution Approach 1:
The patent performs preliminary analysis by calculating good density for each grid before actual production. This pre-computed knowledge about which factor variable ranges lead to good quality outcomes is stored and can be directly applied during manufacturing. By having the optimal factor variable ranges predetermined through grid analysis, the system enables operators to quickly set appropriate parameters without trial-and-error, thereby maintaining both productivity and manufacturing precision.
Data Source
AI summary
A method of the present disclosure includes (a) retrieving from a memory a plurality of measured values of the factor variable, and a label indicating good or bad of the quality corresponding to each of the plurality of measured values, (b) dividing a factor variable space defined by the factor variable into a plurality of grids by equally dividing a range determined by a maximum value and a minimum value of the plurality of measured values for each factor variable, (c) setting a plurality of candidate areas each of which includes one grid or a plurality of adjacent grids, and deriving, for each of the plurality of candidate areas, a good density based on the label associated with the measured value that is within the candidate area, and (d) selecting one of the plurality of candidate areas as the factor variable area, based on the good density.