Density-Based Clustering for Incomplete Data

In real world, missing values exist in a lot of data sets and cause data incompleteness. However, traditional missing value imputation methods are not suitable for density-based clustering and affect the accuracy of clustering results. To solve this problem, this chapter designs a novel density-based clustering model for incomplete data which executes imputation and clustering at the same time and leverages the intermediate clustering results. Evaluation results on 10 synthetic data sets and 5 real-world data sets demonstrate the effectiveness of the proposed model. We provide the research background of this chapter in Sect. 5.1. In Sect. 5.2, we present some basic symbols and briefly introduce the DBSCAN algorithm. In Sect. 5.3, we discuss the process and details of CI clustering. In Sect. 5.4, to overcome some deficiencies of CI clustering, we develop the LI-clustering algorithm. We analyze the experimental results in Sect. 5.5 and draw conclusions in Sect. 5.6.