Rate Forecast about Aircraft Grounded due to Maintenance with Modified Neural Network

Abstract

It was important for the ground crew to forecast the rate of the aircraft grounded due to maintenance. Using the neural network to forecast the index, it depended on the network structure, the algorithm, the training samples quantity and representation ability. The paper ameliorated the network configuration and arithmetic, It constructed the modified network to forecast the value and used the ameliorated method to enlarge its ability during the forecast process. The results showed the method could solve the generalization capability and the sample problems, the forecast results was meaningful.