Research Of Environmental Pollution Sources Data Anomaly Detection Based On Fuzzy Clustering And BP Neural Network

With the rapid development of Chinese economic construction and social productive forces, environmental pollution problems have become increasingly serious. In 2015, Ningxia Hui Autonomous Region environmental protection department held a special meeting. And the department researches and deploys matters of environmental monitoring and associated law enforcement on automatic monitoring. At present, the Environmental supervision departments monitor a wide range with a large number of monitoring data. But detection techniques are still simple mathematical calculations or manual audits. It can’t meet the environmental regulatory authorities to judge and manage the real environment. This project is supported by the the Ningxia Hui Autonomous Region environmental protection science technology research project Research on the development of intelligent data audit and field information management system for pollution sources. The research proposed environmental pollution source data anomaly detection based on fuzzy clustering and BP neural network. It’s a great significance for intelligent construction of environmental monitoring system based on data mining technology. The main contents of this paper is as following:(1) The anomaly detection of pollution source data based on FCM is researched. The data is clustered by membership degree to determine the correlation between data points. It completes the anomaly detection of environmental pollution source data, improves the processing efficiency of environmental pollution in Ningxia.(2) The anomaly correction based on BP neural network algorithm is researched. The Strong nonlinear fitting ability of BP neural network is used to realize anomaly correction. It provides a reference for monitoring the degree of abnormal data that environmental protection departments effectively assess total,supervise,manage.(3) The improved algorithm based on the combination of FCM algorithm and BP neural network to anomaly correction of pollution source is researched with combining the "Z" type reading method. The improved algorithm improves the quality of input samples, and improves the effect of anomaly correction. And the problem that the traditional BP neural network algorithm is easy to fall into local minimum, slow convergence and strong sample dependence is solved.In this paper, The based on fuzzy clustering and BP neural network anomaly detection and anomaly correction of environmental pollution source data improved algorithm based on the combination of FCM algorithm and BP neural network to anomaly correction are proposed. Using simulation experiment, it proves that the algorithm has a good processing effect on anomaly detection and anomaly correction of pollution source data. Compared with the traditional detection technology of pollution source data, it deals with higher accuracy, faster speed, and provides technical support for Ningxia’s scientific and informational environment supervision.