Effective density-based clustering algorithms for incomplete data

Abstract

Density-based clustering is an important category among clustering algorithms. In real applications, manydatasets suffer from incompleteness. Traditional imputation technologies or other techniques for handling missingvalues are not suitable for density-based clustering and decrease clustering result quality. To avoid these problems, we develop a novel density-based clustering approach for incomplete data based on Bayesian theory, which conductsimputation and clustering concurrently and makes use of intermediate clustering results. To avoid the impact oflow-density areas inside non-convex clusters, we introduce a local imputation clustering algorithm, which aims toimpute points to high-density local areas. The performances of the proposed algorithms are evaluated using tensynthetic datasets and five real-world datasets with induced missing values. The experimental results show theeffectiveness of the proposed algorithms.