Correlation-Aware distance measures for data series

Abstract

The field of data series processing has attracted lots of attention thanks to the increased availability of unprecedented amounts of sequential data. These data are then processed and analyzed using a large variety of techniques, most of which are based on the computation of some distance function. In this study, we evaluate the benefits of incorporating into the distance functions correlation measures, which enable us to capture the associations among neighboring values in the sequence. We propose three such measures, inspired by statistical and probabilistic approaches. We analytically and experimentally demonstrate the benefits of the new measures using the 1NN classification task, and discuss the lessons learned.