Predictive modeling of aircraft systems failure using term frequency-inverse document frequency and random forest

For modern aircraft maintenance systems, corrective maintenance is executed by maintenance technicians on ground using the real-time condition monitoring data during the flight. In this paper, a predictive model is proposed to predict faults with high priority in advance by exploring the historical data of aircraft maintenance systems, and preventive maintenance can be carried out based on the prediction results of the model. Prediction of faults with different priorities in this paper are formulated as a binary classification problem. First, counts of different faults occur in past flights are used as raw data, from which term frequency-inverse document frequency is applied to do feature extraction. Next, classification of different faults is modeled by the random forest algorithm and receiver operating characteristics curve is adopted as the performance metrics. For the training dataset, the proposed method achieves true positive rate 100% and false positive rate 0.13%, while for the testing dataset, the proposed method achieves true positive rate 66.67% and false positive rate 0.13%.