Simulation Analysis of the Monitoring Data Classification Algorithm for Wireless Sensor Networks

The problem of data classification for wireless sensor network is researched. The data monitored by wireless sensor network is high redundant,so if traditional BP neural network is used for classification will result in local optimal solution,poor generalization ability,slow convergence speed and low accuracy problem.Therefore data is difficult to be accurately classified.In order to solve the problem immune algorithm is used to optimize BP neural network classification algorithm. Based on the global convergence and individual diversity maintaining mechanism of immune algorithm,the algorithm is used to search globally and optimize BP neural network weights, and then carry out local search with BP algorithm.The simulation results show that,the algorithm can effectively overcome the shortcoming that training classic BP neural network can easily fall into local extreme, so it can accelerate the network convergence speed and the accurate rate of sensor network data classification has been greatly improved.