Shapelet models have attracted a lot of attention from researchers in the time series community, due in particular to its good classification performance. However, such models only inform about the presence / absence of local temporal patterns. Structural information about the localization of these patterns is ignored. In addition, endto-end learning shapelet models tend to generate meaningless shapelets, leading to poorly interpretable models. In this paper, we aim at designing an interpretable shapelet model that takes into account the localization of the shapelets in the time series. Time series are transformed into feature vectors composed of both a distance and a localization information. Then, we design a hierarchical feature selection process using regularization. This process can be tuned to select, for each shapelet, either only its distance information or both distance and localization information. It is hence possible for every selected shapelet to analyze whether only the presence or the presence and the localization contributed to the decision process improving interpretability of the decision. Experiments show that this feature selection process has competitive performance compared to state-of-the-art shapelet-based classifiers, while providing better interpretability.
More and more sensors are used in industrial systems (machines, plants, factories...) to capture energy consumption. All these sensors produce time series data. Abnormal behaviours leading to overconsumption can be detected by experts and represented by sub-sequences in time series, which are patterns. Predictive time series rules are used to detect new occurrences of these patterns as soon as possible. Standard rule discovery algorithms discretize the time series to perform symbolic rule discovery. The discretization requires fine tuning (dilemma between accuracy and understandability of the rules). The first promising proposal of rule discovery algorithm was proposed by Shokoohi et al, which extracts predictive rules from non-discretized data. An important feature of this algorithm is the distance used to compare two subsequences in a time series. Shokoohi et al. propose to use the Euclidean distance to search candidate rules occurrences. However this distance is not adapted for energy consumption data because occurrences of patterns should have different duration. We propose to use more "elastic" distance measures. In this paper we will compare the detection performance of predictive rules based on several variations of Dynamic Time Warping (DTW) and show the superiority of subsequenceDTW.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.