Complexity measures and features for times series classification

Time series classification is a growing problem in different disciplines due to the progressive digitalization of world. The best state-of-the-art algorithms focus on performance, seeking possible results, leaving interpretability at second level, if any. Furthermore, interpretable proposals are far from providing competitive results. In this work, focused time classification, we propose new representation based robust and complete set features. This allows extracting more meaningful information underlying structure develop effective classifiers whose results much easier interpret than current models. proposed feature using traditional vector-based problems, significantly increasing number techniques available for type problem. To evaluate performance our proposal, have used repository UCR, composed 112 datasets. experimental show that through representation, can be obtained which competitive. More specifically, they obtain no statistically significant differences third-best models state-of-the-art. Apart accuracy, proposal able improve model features proposed.