A Knowledge Acquisition Method for Fault Diagnosis of Airborne Equipments Based on Support Vector Regression Machine

Fault diagnosis of airborne equipments is of great significance, while the knowledge for fault diagnosis is hard to acquire. A knowledge acquisition method for fault diagnosis based on support vector regression machine and rough set theory is presented in this work. Due to the redundancy and incompleteness of the original data, we start with applying support vector regression machine to attain support vectors. Then by using Monte Carlo method, we generate new random data around support vectors and build a complete dataset composed of the original data and the new random data. Finally, the rough set method is used to acquire knowledge for fault diagnosis from the complete dataset. The proposed method leads to an increase in accuracy as well as a decrease in uncertainty.