Skip to Main Content
Traditional clustering approaches usually analyze static datasets in which objects are kept unchanged after being processed, but many practical datasets are dynamically modified which means some previously learned patterns have to be updated accordingly. Re-clustering the whole dataset from scratch is not a good choice due to the frequent data modifications and the limited out-of-service time, so the development of incremental clustering approaches is highly desirable. Besides that, propositional clustering algorithms are not suitable for relational datasets because of their quadratic computational complexity. In this paper, we propose an incremental clustering algorithm that requires only one pass of the relational dataset. The utilization of the Representative Objects and the balanced Search Tree greatly accelerate the learning procedure. Experimental results prove the effectiveness of our algorithm.
Date of Conference: 15-19 Dec. 2008