Supporting subseries nearest neighbor search via approximation

INTRODUCTION Searching for nearest neighbors among a large number of time series is an important data mining task in many application domains varying from economics to physics to experimental science in general. Time series are usually modeled as high-dimensional points and proximity of time series is often measured by some metric distance function on their corresponding points [12, 10, 4]. Depending on application needs, time series nearest neighbor searches may be classified into three categories: Whole series query Given a query series and a positive