摘要
通过对自组织数据地图(Self Organizing Map,SOM)理论及技术的研究和应用,力求找到水源水质分析的新方法。以天津市引滦、引黄水源水质数据为例,采用批处理SOM算法为计算核心,利用MATLAB语言编制的SOM工具箱开发了可视化水源水质分析软件包。该软件包通过SOM的训练可对数据特征进行分析和显示,并在此基础上进一步分析SOM训练结果,将监测数据分成5种不同的类型,实现了数据跟踪和新的监测数据的自动归类。应用表明,该方法以多种图形形式直观、综合、深入地分析水源水质,可为水厂水源水质特征的辨识和相应措施的采取提供决策支持。
A new approach for water quality analysis of water source was expected through the study on Self Organizing Map （SOM）. Taking the water quality data from Luanhe River and Huanghe River for example, a set of software for visual water quality analysis was developed based on the batch version algorism of SOM and SOM toolbox in MATLAB environment. This software can analyze and display data characters through the training of SOM. With the analysis from the training results, the data are divided into five categories. Furthermore, it is discovered that the monitored data can be tracked and the new data can be classified automatically. Through application, this method helps to analyze the source water quality in depth by several kinds of graphics, which supplies the decision of recognizing the source water quality characters and taking corresponding measures.
出处
《中国给水排水》
CAS
CSCD
北大核心
2006年第11期60-64,共5页
China Water & Wastewater
基金
国家自然科学基金资助项目(50578108)
天津市科技发展计划资助项目(033113811)
关键词
水源
水质分析
SOM
可视化
water source
water quality analysis
SOM
visualization
