引用本文:陈蓓,徐国伟,王文超,孙小妹,孙培冬,丁彦蕊.PSO和SVM混合算法确定太湖入湖河流水质主要影响因子[J].环境监控与预警,2012,4(2):7-10
.Study on the key factors influenced the water quality of rivers flowing into Taihu Lake using PSO and SVM hybrid algorithm[J].Environmental Monitoring and Forewarning,2012,4(2):7-10
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 2639次   下载 1240 本文二维码信息
码上扫一扫!
分享到: 微信 更多
PSO和SVM混合算法确定太湖入湖河流水质主要影响因子
陈蓓,徐国伟,王文超,孙小妹,孙培冬,丁彦蕊1,2,3
1.无锡市滨湖区环境监测站,江苏 无锡 214072;2.江南大学物联网工程学院,江苏 无锡 214122;3.江南大学化学与材料工程学院,江苏 无锡 214122
摘要:
以影响太湖入湖河流水质的24个因子值为研究对象,将PSO算法与SVM算法相结合。PSO算法用于优化SVM算法的参数c和g,以利于快速、高效地确定c和g的全局最优值;SVM算法基于最优的c和g,分别以24,21,18,15,12,9和6个因子作为特征向量预测水质的污染程度。结果表明,当特征向量为9个影响因子时预测率最高。其参数c=18.56,g=1.35,对应的预测率为:全局预测率92.59%,重度污染水质预测率88.89%,轻度污染水质预测率94.45%。因此,通过PSO和SVM混合算法,可以确定影响太湖入湖河流水质的主要因子,利用这些主要因子对水质进行预测预警,不但可以节省时间,而且可以得到精确的结果。
关键词:  粒子群优化算法  支持向量机  水体水质  影响因素
DOI:
分类号:
基金项目:江苏省环境科研基金项目(0902)
Study on the key factors influenced the water quality of rivers flowing into Taihu Lake using PSO and SVM hybrid algorithm
Abstract:
In this paper, we have studied 24 factors that influenced the water quality of rivers flowing into Taihu Lake by combining the PSO and the SVM algorithm.. The PSO is used to optimize the parameters c and g in SVM, so that the global optimum value of c and g could be searched efficiently and rapidly. Then we use SVM algorithm and take 24,21,18,15,12,9 and 6 influence factors as feature vectors to predict water quality based on the optimal c and g. The results showed that the prediction accuracy is the highest when 9 influence factors is the feature vector. The values of parameter c and g are 18.56 and 1.35 respectively. The corresponding prediction accuracies are computed as follows: the global prediction accuracy is 92.59%, the prediction accuracy of severe pollution water quality is88.89%, the lightly polluted water quality is 94.45%. Therefore, warning prediction of water quality using these factors through the method of PSO and SVM hybrid algorithm is time saving and accurate.
Key words:  Particle swarm optimization  supported vector machines  water quality  impact factor