The prediction of power distribution transformer faults frequency based on high-order Markov

The accurate prediction of power distribution transformer faults frequency would bring certain economic benefits for electronic companies´ production run, in order to avoid outer factors affecting prediction results, high-order Markov is applied to prediction. Firstly we make some preprocessing for the historical series of faults frequency to eliminate their outlier; Then we do the clustering, dividing the historical data into five classifications, next use high-order Markov model to forecast faults frequency and analysis the deviation.