当前位置: 首页 > 文章 > BP神经网络在渭于河流域土壤盐渍化预测中的应用 新疆农业科学 2013,50 (4) 774-779
Position: Home > Articles > Application of BP Network Model for Predicting Soil Secondary Salinization in Weigan River Basin Xinjiang Agricultural Sciences 2013,50 (4) 774-779

BP神经网络在渭于河流域土壤盐渍化预测中的应用

作  者:
买买提·沙吾提;塔西甫拉提·特依拜;丁建丽;张飞
单  位:
新疆大学资源与环境科学学院
关键词:
渭干河流域;BP人工神经网络;土壤盐渍化;多元回归模型
摘  要:
[目的]比较神经网络算法和传统统计建模方法对土壤盐渍化预测模型的效果.[方法]对渭干河流域多年土壤盐渍化和其影响因子进行分析的基础上,采用BP网络的3种算法,建立基于BP神经网络土壤盐渍化预测模型.将预测结果与多元线性回归模型预测结果进行对比分析,讨论线性和非线性方法用于土壤盐渍化预测模型.[结果]与传统的统计建模方法相比BP神经网络结构简单、快捷,预测精度高,很好地再现了土壤盐渍化与其影响因素之间复杂的非线性函数关系;三种BP算法中,基于trainlm算法建立的壤盐渍化预测模型具有较好的推广能力.[结论]BP神经网络的土壤盐渍化预测性能良好,用来可以预测土壤盐渍化情况.
译  名:
Application of BP Network Model for Predicting Soil Secondary Salinization in Weigan River Basin
关键词:
Weigan River Basin%BP Network model%soil salinization%multiple regulation
摘  要:
[ Objective] This project aims to compare the effects of the conventional statistical regulation model and BP network in predicting soil salinization in Weigan River Basin. [ Method] Based on the analysis of the relationship between soil salinity and its influential factors, the effective BP network model was developed for predicting soil salinization in the Weigan River Basin. The model was optimized with three different BP algorithms and the simulation results were compared. Also we have discussed the difference between BP network model and conventional statistical methods for estimation of soil salinization. [ Result] The results showed that the BP network model was simple, quick with higher prediction precision compared with the conventional multiple regulation model. Among the three BP algorithm for the soil salinization forecast, the trainlm algorithm is the most effective. [ Conclusion ] The result presumed that the BP algorithm is useful for the soil salinization forecast.

相似文章

计量
文章访问数: 7
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊