Outlier Detection using Hybrid Approach for Mixed datasets

Abstract

Several approaches of outlier detection are employed in many study areas amongst which distance based and density based outlier detection techniques have gathered most attention of researchers .So we are using hybrid of these two methods. The proposed model uses hybrid of distance and density outlier detection methods and weighted squeezer method for clustering. Advantages of both outlier methods will be combined giving higher result. The clustering algorithm which does not require to specify number of clusters as input which is drawback of many clustering algorithms. Most of the models deals with only single datatype datasets. Here the project deals with mixed datatype datasets. Here we will compare hybrid system with single method system. From performance measures it will be cleared how hybrid system gives better results as compared to single method.
DOI: 10.17762/ijritcc2321-8169.1506103