铁路行车事故预测方法分析与比较

对铁路行车事故的特点和类型进行分析;根据美国铁路2005年安全年报提供的数据,运用灰色系统理论和BP神经网络方法建立铁路行车事故的预测模型;利用MATLAB软件进行预测仿真,比较和分析两种预测方法的精度及特点。结果表明:灰色系统理论预测结果固定,短期效果比较好;BP神经网络预测具有适应性和灵活性,适用于长期预测。采用灰色系统理论和BP神经网络进行铁路行车事故的预测,克服了传统数学统计预测方法中建立复杂的数学模型,预测准确性低的缺点,对预防和控制铁路事故的发生,降低事故损失具有现实意义。

Analysis and Comparison of Predication Methods for Railway Train Accidents

The characteristics and types of the railway train accidents are analyzed.According to the data of America railway safe Annual Report in 2005,gray theory and BP neural network method are applied to build railway train accident predication model,and the MATLAB software is used to simulate,the predication accuracy and characteristics of two methods are compared and analyzed.The results show that the gray theory has a fixed predication outcome and better short-term effect,while BP neural network has a better long-term effect due to its adaptability and flexibility.Gray theory and BP neural network predication methods overcome the shortcomings of low accuracy of traditional mathematical statistical predication methods as well as creating complex mathematical models.This study is of significant meaning to the prevention and control of railway accidents and loss reduction.