The Mutual Information based correlation analysis between fault types and monitor data

The fault monitor data analysis of aviation equipment is essential for the failure diagnosis and further researches. Here we propose a new approach based on Mutual Information to measure the correlation of the fault types and the data indexes, particularly in how to analysis the correlation of discrete variables and continuous variables. Then the fire control system of aircraft is taken for example to make sure the fault symptom class. Furthermore, we compare the results with that by Correlation Coefficient method and draw the conclusion that the Mutual Information method based on entropy makes more sense.