Application of Neural Network in Ranking Police Alarms

It is a practical need to handle police alarms reported effectively to realize automatic ranking. In this paper, we proposed a police alarm-reporting classification framework involving rough set attribute reduction, genetic algorithm and back propagation neural network. In this framework, we first obtain the reduction of redundant attribute with the theory of rough set, then combine the genetic algorithm and back propagation algorithm to train the network model. A case study on classification of public alarm rank demonstrates the usability of the framework.