A statistical comparison of neuroclassifiers and feature selection methods for gearbox fault diagnosis under realistic conditions

Engineering & Materials Science

• Failure analysis 100%
• Feature extraction 89%
• Neural networks 70%
• Statistical tests 33%
• Unsupervised learning 33%
• Factorization 28%
• Power transmission 28%
• Entropy 24%
• Learning algorithms 24%
• Genetic algorithms 20%
• Industry 12%
• Costs 11%

Medicine & Life Sciences

• Learning 85%
• Computer Neural Networks 50%
• Entropy 46%
• Equipment and Supplies 46%
• Industry 35%
• ROC Curve 33%
• Area Under Curve 31%
• Costs and Cost Analysis 24%