An Algorithm based on piecewise slope transformation distance for short time series similarity measure

摘要

Aiming at the irregular and uneven feature of medical time series data, a novel algorithm based on piecewise slope transformation distance for short time series similarity measure is propose. We firstly do some preprocess based on algorithm for key points selected, make the data curve to zigzag shape, then, we measure the distance between two curves based on piecewise slope transformation algorithm. By experiments, conclusion can be draw that this new approach can measure distance rapidly and correctly, especially appropriate to short time series data.