宜宾市酸雨pH值预测的偏最小二乘回归分析

酸雨pH值受酸性离子(有机酸、无机酸)和碱性离子的影响。这些影响因素之间存在多重相关性。用一般最小二乘回归法建模预测pH值,估计参数存在着很大的误差,而且物理意义明显不足。本文以宜宾市区2002-2003年的27组降雨监测数据作为样本数据,应用偏最小二乘回归技术建立pH值预测模型,克服了自变量之间的多重相关性的问题。与最小二乘回归法相比更具有先进性,计算结果更为可靠;在确定了模型可行性后,分析比较了影响宜宾市区酸雨pH值的离子的重要性和离子来源。

Partial Least---a squares Regressive Model and its Application to the Analysis of pH Values in Yibin's Acid Precipitation

pH value of acid precipitation is affected by acid ions (organic acid and inorganic acid ) and alkaline ions. There are some multiple correlations between these factors. As the multiple correlations are existed, the estimated regressive parameters with the least squares method include a good deal of errors in the multiple regressive equations, which cannot reflect its physical meaning. The paper apply partial least-squares regressive method to the analysis of pH values in Yibin's acid precipitation. 27 group dates of 2002-2003 are used in constructing the PLS model. The partial least-square method can easily solve the multiple correlation problems. It is more scientific and its result is more trustiness. After make sure the feasibility of the PLS model, the ion, which affects the pH of Yibin's acid precipitation more, has been discussed.