Prediction Model of Civil Aviation Incidents Based on Grey Neural Network
Hui Li · Zhongguo anquan kexue xuebao · 2012
The corresponding safety measures were made to prevent civil aviation accidents through analysis and forecasts of civil aviation incidents to grasp the trend of development of civil aviation incidents.Based on the grey prediction model,the ways to apply neural network to the forecasting practice were presented,and the characteristics of civil aviation incidents were analyzed.The serial grey neural network combination prediction model of civil aviation incidents was advanced.Firstly,the grey method was applied to weaken the fluctuations in data sequence,and then the model was built taking advantage of the ability of processing non-linear information of neural network.The gray prediction model,neural network model and the gray neural network model were tested according to the data about civil aviation incidents in the recent decade.The numbers of civil aviation incidents for the next three years were predicted.The Matlab software was used for simulating,and the prediction accuracy and characteristics of the three methods were analyzed and compared.The results show that the accuracy of the grey neural network prediction model is higher than that of the single grey prediction model or neural network model.The grey neural network prediction model requires less samples,involves only simple calculations,and can be realized easily.