A Novel Density based Clustering Method using Nearest and Farthest Neighbor with PCA

Common nearest-neighbor density estimators usually do not work well for high dimensional datasets. Moreover, they have high time complexity of O(n2) and require high memory usage especially when indexing is used. In order to overcome these limitations, we proposed a new method that calculates distances to nearest and farthest neighbor nodes to create dataset subgroups. Therefore computational time complexity becomes of O(nlogn) and space complexity becomes constant. After subgroup formation, assembling technique is used to derive correct clusters. In order to overcome high dimensional datasets problem, Principal Component Analysis (PCA) in the clustering method is used, which preprocesses high-dimensional data. Many experiments on synthetic data sets are carried out to demonstrate the feasibility of the proposed method. Furthermore we compared this algorithm to the similar algorithm –DBSCAN- on real-world datasets and the results showed significantly higher accuracy of the proposed method.