ICA: An Incremental Clustering Algorithm Based on OPTICS

Clustering algorithms play an important role in data mining no matter whether they are used as a stand-alone tool or as a preprocessing step for further analysis on the data. With the arrival of the information era, the speed of data generation is faster and faster. As a result, clustering algorithms, such as OPTICS, that can only be operated on the static dataset can’t meet the new requirements. Motivated by the demand of clustering analysis on dynamic datasets efficiently, in this paper, we propose an incremental clustering algorithm (ICA) based on OPTICS. The result of ICA is a cluster-ordering structure which is some similar to the result of OPTICS. In ICA, we delete the parameters ɛ and MinPts that should be preset by users in OPTICS and reachability-distance is also replaced by Distance which is easier to compute and understand. As a result, ICA is much more efficient compared with OPTICS. In addition, we propose a method named automatically extract technique to extract the clusters from the cluster-ordering structure based on the users’ needs. Our performance evaluation through a series of experiments demonstrates the effectiveness and efficiency of our algorithm. Specially, we present a detailed comparison of ICA and OPTICS and the results illustrate that ICA is much more suitable for clustering the dynamic datasets, i.e., some new data objects are added into the datasets as time goes on.