Clustering human perception of environment impact using Rough Set Theory

Abstract

Rough set is a set theory which is have been applied in the many areas. One of them is in data mining. The utilization of feature selection and clustering methods, that are a part of data mining application, could contribute for decision support. This paper investigates the application of rough set theory to select attribute and cluster environment impact. The Maximum Dependency Attribute (MDA) and fuzzy partition based on indiscernible relation are used to select the most important impact and cluster the object using the selected attributes, respectively. The data are collected from the field survey at identifying the environmental impact experienced by several communities in Yogyakarta, Indonesia. The results show that the water quality is the important attribute on physical and chemical aspects. Furthermore, on economic aspect, the highest attributes are immigration and employee absorption. Moreover, the number of cluster recommended is 9 based on the silhouette coefficient which is rising 0.9. This paper can be used to make recommendation to improve the quality of social environment.