Plasma Processing State Prediction With Ranked Feature Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for abnormality detection in plasma processing apparatuses face challenges with long calculation times and reduced reliability due to the inclusion of unnecessary features, which can impair the detection of main information for abnormality detection.
Innovation Solution
A state prediction apparatus that determines a first set of features indicating a normal state, calculates a second set of features, and generates a model using a subset of features selected in descending order, allowing for reliable abnormality detection in short calculation time without excluding main features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If principal component analysis is performed using a large number of extracted features, then main information can be comprehensively captured, but calculation time becomes excessively long
Solution Approach 1:
The patent extracts and removes unnecessary features from the feature set before performing principal component analysis. By identifying and eliminating features that do not contribute meaningfully to abnormality detection, the system reduces the input dimensionality while preserving the essential information needed for reliable detection, thus shortening calculation time without sacrificing detection reliability.
Solution Approach 2:
The patent segments the feature selection process into multiple stages: initial feature extraction, unnecessary feature removal, and then principal component analysis on the refined feature set. This segmented approach allows the system to systematically reduce feature dimensionality while maintaining the quality of information used for abnormality detection.
2Quantity of substance
If unnecessary features are included in the analysis, then comprehensive data coverage is achieved, but detection reliability deteriorates due to noise
Solution Approach 1:
The patent applies feature removal techniques to extract and eliminate unnecessary features from the dataset. By identifying features that represent noise rather than meaningful signal, the system removes these detrimental elements while retaining the comprehensive coverage of important parameters, thereby improving detection reliability.
Solution Approach 2:
The patent changes the parameter composition of the feature set by dynamically adjusting which features are included in the analysis. Through parameter optimization and feature selection, the system transforms the feature set from a static comprehensive collection to a dynamic optimized subset that maximizes detection reliability while minimizing noise impact.
3Productivity
If feature selection is performed to reduce calculation time, then processing speed improves, but main features for abnormality detection may be excluded
Solution Approach 1:
The patent performs preliminary feature evaluation and removal before conducting principal component analysis. By预先 identifying and removing unnecessary features in advance, the system prepares an optimized feature set that ensures main detection features are preserved while reducing computational load, thus achieving both high processing speed and maintained reliability.
Solution Approach 2:
The patent employs feedback mechanisms in the feature selection process, where the system evaluates the impact of feature removal on detection performance and adjusts the feature set accordingly. This feedback-driven approach ensures that feature reduction does not compromise the inclusion of critical features needed for accurate abnormality detection.
Data Source
AI summary
Provided is a state prediction apparatus that predicts a state of the plasma processing apparatus, a first set of features that indicates the state of the plasma processing apparatus is determined based on monitored data of the plasma processing apparatus in a normal state, a second set of features that indicates the state of the plasma processing apparatus is determined based on monitored data of the plasma processing apparatus, the features in the second set are calculated by using the features in the first set, a model that predicts the state of the plasma processing apparatus is generated by using a subset of the first set of features, which is composed of the same kind of features selected in descending order of the calculated features in the second set, and the state of the plasma processing apparatus is predicted by using the generated model.