Fault rate prediction method of ATP (Automatic Train Protection) vehicle-mounted equipment

The invention discloses a fault rate prediction method of ATP (Automatic Train Protection) vehicle-mounted equipment. The fault rate prediction method comprises the following steps: obtaining the historical data of an ATP vehicle-mounted equipment operation fault; reconstructing the phase space of the obtained historical data of the ATP vehicle-mounted equipment operation fault to high-dimensionality spatial data; on the basis of the reconstructed high-dimensionality spatial data, combining with a Lyapunov index method to judge whether the historical data of the ATP vehicle-mounted equipment operation fault has chaotic characteristics or not; and if the historical data of the ATP vehicle-mounted equipment operation fault has the chaotic characteristics, establishing a chaotic prediction model, adopting a SVM (Support Vector Machine) to fit a relationship between the prediction data and the historical data of the ATP vehicle-mounted equipment operation fault so as to predict a dynamic change tendency of an ATP equipment fault rate. The method disclosed by the invention can predict the tendency of the ATP equipment fault rate in short time so as to provide scientific references for reasonably carrying the optimal configuration of ATP spare parts.