A New Patterns Recognition Method Based on Fuzzy Rough Sets
Guang Tian, Xin Jie Shao, Jin Hua Liu, Yun Guang Qi · Applied Mechanics and Materials · 2013
Gearbox is the key component of self-propelled gun, and it is prone to failure. So, its performance will directly affect the fight maneuverability of self-propelled gun. To detect the running state of gearbox and find incipient fault in time, has very important significance.This paper put forward a new patterns recognition method based on fuzzy rough sets, which can reduce the information missing during the discretization process of fuzzy rough, and overcome the disadvantage of artificially determining class number through F test, applied the fuzzy decision table to attributes reduction, got the clear and concise pattern rules. The gearbox fault patterns recognition result showed that the propsed method can greatly enhance the patterns recognition accuracy, and it is a good method in patterns recognition application.