Prediction Model Feature Grouping for Low-Frequency Machine Failures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately predict machine failures in environments with multiple failure modes and operation modes, particularly when failure frequencies are low, due to the complexity of feature identification and the need for manual preparation of features, without adequately addressing the variety of failure modes and operation modes.
Innovation Solution
A prediction model construction support system that divides explanatory variables into groups to improve prediction accuracy by calculating a score based on the support ratio and confidence of identification features within each group, using a model construction support system to facilitate the search for appropriate features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual feature identification and selection methods are used, then the prediction model can be constructed, but the workload becomes large and the process becomes complicated due to the need for complicated progress management for various factor analyses
Solution Approach 1:
The system automatically performs feature identification, group division, and score calculation without requiring manual intervention. The computer executes the prediction model construction support program to autonomously divide explanatory variables into groups, calculate identification feature scores, and determine optimal group divisions, eliminating the need for manual progress management and factor analysis while maintaining high prediction accuracy
Solution Approach 2:
The patent replaces manual mechanical processes of feature selection and group division with automated computational processes. The system uses algorithms to automatically divide explanatory variables into groups, calculate scores based on support ratios and identification features, and determine optimal divisions, substituting human analysts' manual work with automated mechanical computation
2Measurement precision
If existing identification methods such as decision tree, random forest, and XGBoost are used, then the feature selection can be performed, but it becomes difficult to ascertain a sign for failure when the frequency of machine failure is low
Solution Approach 1:
The patent divides the set of explanatory variables into multiple groups based on identification features, and performs separate score calculations for each group. This segmentation allows the system to focus on specific subsets of variables that are most relevant to failure prediction, improving the ability to detect failure signs even when failure frequency is low by concentrating analysis on the most predictive variable groups
Solution Approach 2:
The system changes the approach from traditional single-model identification methods to a group-based scoring system that calculates support ratios and identification feature scores. By transforming the problem into a multi-group scoring framework, the system can better handle low-frequency failure events through systematic evaluation of multiple variable groups with different characteristics
3Extent of automation
If search methods such as AutoML, genetic algorithm, and reinforcement learning are used, then the feature search can be automated, but it is necessary for a person to prepare the feature in advance
Solution Approach 1:
The system performs preliminary automatic division of explanatory variables into groups and calculation of identification features before the actual prediction model construction. By pre-processing the data through automated group division and score calculation, the system eliminates the need for manual feature preparation while maintaining full automation throughout the process
Solution Approach 2:
The patent creates a universal system that can handle various types of explanatory variables and failure modes through a standardized group division and scoring approach. The system's multi-functional capability allows it to automatically process different datasets and prediction tasks without requiring task-specific feature preparation, achieving both automation and ease of use
Data Source
AI summary
A model construction support system supports searching for a feature used to construct a prediction model that outputs an objective variable related to a predicted event for a machine based on explanatory variables, and a division method for dividing the explanatory variables into groups to improve calculation accuracy of the objective variables based on the prediction model. The system divides the explanatory variables into a plurality of groups, calculates accuracy of the features set based on the explanatory variable in the groups, and calculates a score of the feature in the groups based on the accuracy and a support ratio of the explanatory variable to all of the explanatory variables before division. The system calculates accuracy of a group division feature used to divide the explanatory variables, and a score in the groups based on the score and the accuracy in the groups.