Establishment and Application of BP Neural Network Prediction Model of Xinxue River Constructed Wetland, China Based on Data Expansion

摘要

Over the course of study the BPNN prediction model of Xinxue River constructed wetland, model boundary was found as an important restrict factor of model accuracy. In the paper data expansion was selected to solve this problem, and the extended data was used to re-built the prediction model, results showed that the performance of model was effectively improved.