基于投影寻踪的集对分析模型及其在地下水质评价中的应用  

Application of projection pursuit based set pair analysis model in the evaluation of underground water quality

作　　者：李陶[1] 付强[1] 丁红[1] 

机构地区：[1]东北农业大学水利与建筑学院,哈尔滨150030

出　　处：《东北农业大学学报》2010年第4期48-52,共5页Journal of Northeast Agricultural University

基　　金：黑龙江省科技攻关项目(GB06B106-7)

摘　　要：文章针对传统评价方法不能很好地解决评价因子与水质等级间复杂的非线性关系,以及主观确定权重等问题,提出基于投影寻踪的改进集对分析模型(RAGAPP-SPA)。利用最佳投影方向作为指标权重,并通过完善Ⅴ级相邻级别联系度为-1的不足,假设存在Ⅵ级的指标值,从而得出联系度,进而对涡北井田地下水质进行了等级评价。结果表明,13个地区评价结果基本与原方法相同,权重符合实际规律。基于投影寻踪的改进型集对分析,能够客观确定权重,改进后的集对分析模型等级划分更细致,对检测数据利用率高,可应用于各种非线性、非正态高维数据评价研究中。A new method, improved Set Pair Analysis Model based on Projection Pursuit （RAGAPP-SPA）, was used to solve the complicated non-linear relation between the evaluation factors and water quality, and issues such as subjective weight that the traditional methods can not do well. The relational degree can be obtained by using the optimum projection vector as the index weight, perfecting the deficiency that adjacent relational degree of V degree is -1, and assuming the existence of six degree indices, and then evaluated quality of underground water of minefield named Wobei. The evaluation results for 13 regions proved the same as the actual method and the weight accords with the actual law. By utilizing the RAGAPP-SPA, the weight could be determined objectively, the grade could be divided carefully, the tested data could be utilized greater, and can be applied to research a variety of non-linear and non-normal high-dimensional data.

关 键 词：集对分析 权重 投影寻踪 联系度 评价水质 

分 类 号：X824[环境科学与工程—环境工程]