Root Cause Prediction for Failures in Semiconductor Industry: A Genetic Algorithm–ML Approach
Rammal, A., et al.
Scientific Reports (2023)
Abstract
This paper presents a hybrid approach combining genetic algorithms with machine learning
for accurate root cause prediction in semiconductor manufacturing. Our methodology
leverages the optimization capabilities of genetic algorithms to enhance the performance
of machine learning models in identifying failure root causes.
Key Contributions
Novel hybrid approach combining genetic algorithms with machine learning
Advanced feature selection and optimization techniques
Improved accuracy in root cause identification for complex failure scenarios
Scalable framework for processing large-scale manufacturing data
Real-time prediction capabilities for proactive failure prevention
Key Findings
Achieved 88% accuracy in root cause prediction compared to traditional methods
Reduced false positive rate by 45% through genetic algorithm optimization
Improved prediction speed by 3x through efficient algorithm design
Enhanced ability to handle multi-dimensional failure data
Significant improvement in early failure detection capabilities
Methodology
Our approach employs genetic algorithms for feature selection and hyperparameter
optimization, combined with advanced machine learning models for pattern recognition.
The hybrid system processes multi-dimensional manufacturing data to identify
complex failure patterns and predict root causes with high accuracy.