A graph model for preventing railway accidents based on the maximal information coefficient

A number of factors influences railway safety. It is an important work to identify im portant influencing factors and to build the relationship between railway accident its influencing factors. The maximal information coefficient (MIC) is a good measure of dependence for two-variable relationships which can capture a wide range of associactions. Employing MIC, a graph model is proposed for preventing railway accidents which avoids complex mathematical computation. In the graph, nodes denote influencing factors of railway accidents and edges represent dependence of the two linked factors. With the increasing of dependence level, the graph changes from a globally coupled graph to isolated points. Moreover, the important influencing factors are identified from many o factors which are the monitor key. Then the relationship between railway accident and ti important influencing factors is obtained by employing the artificial neural networks. With the relationship, a warning mechanism is built by giving the dangerous zone. If the related factors fall into the dangerous zone in railway operations, the warning level should be raised. The built warning mechanism can prevent railway accidents and can promote railway safety.