Clustering is indispensable to obtain a general view of series data from a number of data such as gene expression profiles. We propose a novel metric for clustering. The proposed metric automatically normalizes data to minimize a logarithmic scale distance between the data series.
