情報処理学会 第75回全国大会講演要旨

Density-based clustering algorithms, such as DBSCAN, usually have a difficulty in selecting appropriate parameters. Recently, the FN-DBSCAN algorithm extended density-based clustering algorithms with the fuzzy set theory and solved this problem. However, FN-DBSCAN has a time complexity of o(n^2) which indicates that it is not suitable to deal with large scale of data. In this study, we propose a novel clustering algorithm called landmark FN-DBSCAN which has a linear time and space complexity to the size of input data and provides a good quality of clustering.