A model-based clustering for time-series with irregular interval

Clustering problems are central to many knowledge discovery and data mining tasks. However, most existing clustering methods can only work with fixed-interval representations of data patterns, ignoring the variance of time axis. In this paper, we study the clustering of data patterns that are sample in irregular interval. In the paper, a model-based approach using cepstrum distance metrics and autoregressive conditional duration (ACD) model is proposed. Experimental results on real datasets show that this method is generally effective in clustering irregular space time series, and conclusion inferred from experimental results agrees with the market microstructure theories.