Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction

A wide range of applications based on sequential data, named time series, have become increasingly popular in recent years, mainly those based on the Internet of Things (IoT). Several different machine learning algorithms exploit the patterns extracted from sequential data to support multiple tasks....

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Ricardo Petri Silva, Bruno Bogaz Zarpelão, Alberto Cano, Sylvio Barbon Junior
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	time series segmentation stationarity analysis time series prediction improvement size reduction in time series Chemical technology TP1-1185
Acceso en línea:	https://doaj.org/article/35fab8c149124520ace8803a6f6d67c6
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:35fab8c149124520ace8803a6f6d67c6
record_format	dspace
spelling	oai:doaj.org-article:35fab8c149124520ace8803a6f6d67c62021-11-11T19:16:41ZTime Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction10.3390/s212173331424-8220https://doaj.org/article/35fab8c149124520ace8803a6f6d67c62021-11-01T00:00:00Zhttps://www.mdpi.com/1424-8220/21/21/7333https://doaj.org/toc/1424-8220A wide range of applications based on sequential data, named time series, have become increasingly popular in recent years, mainly those based on the Internet of Things (IoT). Several different machine learning algorithms exploit the patterns extracted from sequential data to support multiple tasks. However, this data can suffer from unreliable readings that can lead to low accuracy models due to the low-quality training sets available. Detecting the change point between high representative segments is an important ally to find and thread biased subsequences. By constructing a framework based on the Augmented Dickey-Fuller (ADF) test for data stationarity, two proposals to automatically segment subsequences in a time series were developed. The former proposal, called Change Detector segmentation, relies on change detection methods of data stream mining. The latter, called ADF-based segmentation, is constructed on a new change detector derived from the ADF test only. Experiments over real-file IoT databases and benchmarks showed the improvement provided by our proposals for prediction tasks with traditional Autoregressive integrated moving average (ARIMA) and Deep Learning (Long short-term memory and Temporal Convolutional Networks) methods. Results obtained by the Long short-term memory predictive model reduced the relative prediction error from 1 to 0.67, compared to time series without segmentation.Ricardo Petri SilvaBruno Bogaz ZarpelãoAlberto CanoSylvio Barbon JuniorMDPI AGarticletime series segmentationstationarity analysistime series prediction improvementsize reduction in time seriesChemical technologyTP1-1185ENSensors, Vol 21, Iss 7333, p 7333 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	time series segmentation stationarity analysis time series prediction improvement size reduction in time series Chemical technology TP1-1185
spellingShingle	time series segmentation stationarity analysis time series prediction improvement size reduction in time series Chemical technology TP1-1185 Ricardo Petri Silva Bruno Bogaz Zarpelão Alberto Cano Sylvio Barbon Junior Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction
description	A wide range of applications based on sequential data, named time series, have become increasingly popular in recent years, mainly those based on the Internet of Things (IoT). Several different machine learning algorithms exploit the patterns extracted from sequential data to support multiple tasks. However, this data can suffer from unreliable readings that can lead to low accuracy models due to the low-quality training sets available. Detecting the change point between high representative segments is an important ally to find and thread biased subsequences. By constructing a framework based on the Augmented Dickey-Fuller (ADF) test for data stationarity, two proposals to automatically segment subsequences in a time series were developed. The former proposal, called Change Detector segmentation, relies on change detection methods of data stream mining. The latter, called ADF-based segmentation, is constructed on a new change detector derived from the ADF test only. Experiments over real-file IoT databases and benchmarks showed the improvement provided by our proposals for prediction tasks with traditional Autoregressive integrated moving average (ARIMA) and Deep Learning (Long short-term memory and Temporal Convolutional Networks) methods. Results obtained by the Long short-term memory predictive model reduced the relative prediction error from 1 to 0.67, compared to time series without segmentation.
format	article
author	Ricardo Petri Silva Bruno Bogaz Zarpelão Alberto Cano Sylvio Barbon Junior
author_facet	Ricardo Petri Silva Bruno Bogaz Zarpelão Alberto Cano Sylvio Barbon Junior
author_sort	Ricardo Petri Silva
title	Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction
title_short	Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction
title_full	Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction
title_fullStr	Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction
title_full_unstemmed	Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction
title_sort	time series segmentation based on stationarity analysis to improve new samples prediction
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/35fab8c149124520ace8803a6f6d67c6
work_keys_str_mv	AT ricardopetrisilva timeseriessegmentationbasedonstationarityanalysistoimprovenewsamplesprediction AT brunobogazzarpelao timeseriessegmentationbasedonstationarityanalysistoimprovenewsamplesprediction AT albertocano timeseriessegmentationbasedonstationarityanalysistoimprovenewsamplesprediction AT sylviobarbonjunior timeseriessegmentationbasedonstationarityanalysistoimprovenewsamplesprediction
_version_	1718431576845451264

Time Series Segmentation Based on Stationarity Analysis to Improve New Samples Prediction

Ejemplares similares