THE RESEARCH OF RAINFALL PREDICTION MODELS BASED ON MATLAB NEURAL NETWORK

摘要：

The continuously cloudy or rainy forecast is an important basis that is used to make choice of wheat harvest time but multiple regression weather forecast models hardly content the rate of required accuracy. Matlab neural network toolbox is composed of a series of typical neural network activation functions that make computing network output into calling activation functions. BP artificial neural network that is based on Matlab platform and utilizes error back propagation algorithm to revise network weight has dynamic frame characteristics and is convenient for constructing network and programming. After it has been trained by input forecast samples, network forecast model that has three neural cells possesses very good generalization capability. After we contrast fitting rate and accuracy rate of network model with ones of regression model, network model has a distinct advantage over regression model.

关键词： forecast models BP network learning algorithm network training

作者: Xianggen Gan Lihong Chen Dongbao Yang Guang Liu

作者单位: Jiangxi Vocational &Technical College of Information Application, Nanchang, China Nanchang Meteorological Bureau, Nanchang, China Xinyu Meteorological Bureau, Xinyu, China

会议类型: 国际会议

会议名称: 2011 IEEE International Conference on Cloud Computing and Intelligence Systems(2011年第一届IEEE云计算与智能系统国际会议 IEEE CCIS2011)

会议地点: 北京

会议语种:英文

页码: 45-48

在线出版日期: 2011-09-15（万方平台首次上网日期，不代表论文的发表时间）