A Clustering Approach to Distinguish the Change-Point

In Phase I, it is possible that the baseline may represent more than one distribution. This is the problem when we use control charts to detect changes in the pattern of data over time. As we known, the change-point estimation problem is to identify the real time of the change. This paper proposes a new method to distinguish the change-point. The proposed method is based on the clustering techniques, a moving window theory, and the probability density theory. Compared with classical and robust estimation procedures, simulation studies show that our method is usually better and sometimes much better at distinguishing the change-point.