Bearing fault classification using firefly clustering

This paper presents a novel bearing fault classification method based on firefly algorithm. This method is originated by nature-inspired swarm intelligence, which contributes to a supervised classification task by labeling a few samples. The main advantages of this method are: 1) first, it does not require a large amount of samples in the population; 2) second, the coding of firefly is very simple; 3) third, the convergence rate is high. Simulation on Iris data clustering and experiments on bearing fault classification validate the effectiveness of the proposed method.