基于最优分割理论的酸雨监控点位的分类研究

Classification of Acid Rain Monitoring Sites Based on Fisher Method

Fisher最优分割法是用离差平方和来表示同类样本之间的差异程度,通过简便的计算步骤和作图,确定最优分类数,使同类样本间的差异最小,各类别样本间的差异最大,并用F检验法检验最优分类数的合理性。文章根据Fisher最优分割法的理论内涵,研究其应用于酸雨环境监测优化布点的可行性。以漳州市沿海区域监测点为研究实例,用Fisher最优分割法确定最优的分类情况和最优的优化布点方案。结果表明,原布设的9个监测点用Fisher最优分割法优化为3个典型点位,可客观地反映漳州市酸雨控制区的整体污染状况和污染趋势,该方法理论内涵直观、计算简便、重复性好,对环境质量监测有重要的意义。

Fisher optimal dissection method was applied to determine the optimum number of ordinal environmental monitoring stations. Error functions were introduced into the objective methodology to make the discrepancies maximization in different categories and minimization in the same category. The most desirable number of monitoring sites was identified through the change trend of minimum error functions and the rationality of the above classification results was tested by F detected method. The acid rain monitoring points of Zhangzhou was presented as an empirical research example to explore the application possibility of optimal stations selections in acid rain environmental quality monitoring based on the Fisher basic theory. Results showed that original 9 monitoring points could be optimized into three typical sites. Connotation of Fisher method which is intuitive and clearly is essential to increase the clustering effect of optimized layout for monitoring points. The Fisher optimal dissection methodology could be a useful decision support tool for expanding the application areas in other environment monitoring optimal stations selection.