Self-learning of multivariate time series using perceptually important points

Research output: Contribution to journalArticleScientificpeer-review

Abstract

In machine learning, positive-unlabelled PU learning is a special case within semi-supervised learning. In positive-unlabelled learning, the training set contains some positive examples and a set of unlabelled examples from both the positive and negative classes. Positive-unlabelled learning has gained attention in many domains, especially in time-series data, in which the obtainment of labelled data is challenging. Examples which originate from the negative class are especially difficult to acquire. Self-learning is a semi-supervised method capable of PU learning in time-series data. In the self-learning approach, observations are individually added from the unlabelled data into the positive class until a stopping criterion is reached. The model is retrained after each addition with the existent labels. The main problem in self-learning is to know when to stop the learning. There are multiple, different stopping criteria in the literature, but they tend to be inaccurate or challenging to apply. This publication proposes a novel stopping criterion, which is called Peak evaluation using perceptually important points, to address this problem for time-series data. Peak evaluation using perceptually important points is exceptional, as it does not have tunable hyperparameters, which makes it easily applicable to an unsupervised setting. Simultaneously, it is flexible as it does not make any assumptions on the balance of the dataset between the positive and the negative class.

Original languageEnglish
Article number8894745
Pages (from-to)1318-1331
Number of pages14
JournalIEEE/CAA Journal of Automatica Sinica
Volume6
Issue number6
DOIs
Publication statusPublished - Nov 2019
MoE publication typeA1 Journal article-refereed

Fingerprint

Time series
Supervised learning
Learning systems
Labels

Cite this

@article{72ef34ac28494f3897b9b4d9fa7ccd9e,
title = "Self-learning of multivariate time series using perceptually important points",
abstract = "In machine learning, positive-unlabelled PU learning is a special case within semi-supervised learning. In positive-unlabelled learning, the training set contains some positive examples and a set of unlabelled examples from both the positive and negative classes. Positive-unlabelled learning has gained attention in many domains, especially in time-series data, in which the obtainment of labelled data is challenging. Examples which originate from the negative class are especially difficult to acquire. Self-learning is a semi-supervised method capable of PU learning in time-series data. In the self-learning approach, observations are individually added from the unlabelled data into the positive class until a stopping criterion is reached. The model is retrained after each addition with the existent labels. The main problem in self-learning is to know when to stop the learning. There are multiple, different stopping criteria in the literature, but they tend to be inaccurate or challenging to apply. This publication proposes a novel stopping criterion, which is called Peak evaluation using perceptually important points, to address this problem for time-series data. Peak evaluation using perceptually important points is exceptional, as it does not have tunable hyperparameters, which makes it easily applicable to an unsupervised setting. Simultaneously, it is flexible as it does not make any assumptions on the balance of the dataset between the positive and the negative class.",
author = "Timo Lintonen and Tomi Raty",
year = "2019",
month = "11",
doi = "10.1109/JAS.2019.1911777",
language = "English",
volume = "6",
pages = "1318--1331",
journal = "IEEE/CAA Journal of Automatica Sinica",
issn = "2329-9266",
publisher = "IEEE Institute of Electrical and Electronic Engineers",
number = "6",

}

Self-learning of multivariate time series using perceptually important points. / Lintonen, Timo; Raty, Tomi.

In: IEEE/CAA Journal of Automatica Sinica, Vol. 6, No. 6, 8894745, 11.2019, p. 1318-1331.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - Self-learning of multivariate time series using perceptually important points

AU - Lintonen, Timo

AU - Raty, Tomi

PY - 2019/11

Y1 - 2019/11

N2 - In machine learning, positive-unlabelled PU learning is a special case within semi-supervised learning. In positive-unlabelled learning, the training set contains some positive examples and a set of unlabelled examples from both the positive and negative classes. Positive-unlabelled learning has gained attention in many domains, especially in time-series data, in which the obtainment of labelled data is challenging. Examples which originate from the negative class are especially difficult to acquire. Self-learning is a semi-supervised method capable of PU learning in time-series data. In the self-learning approach, observations are individually added from the unlabelled data into the positive class until a stopping criterion is reached. The model is retrained after each addition with the existent labels. The main problem in self-learning is to know when to stop the learning. There are multiple, different stopping criteria in the literature, but they tend to be inaccurate or challenging to apply. This publication proposes a novel stopping criterion, which is called Peak evaluation using perceptually important points, to address this problem for time-series data. Peak evaluation using perceptually important points is exceptional, as it does not have tunable hyperparameters, which makes it easily applicable to an unsupervised setting. Simultaneously, it is flexible as it does not make any assumptions on the balance of the dataset between the positive and the negative class.

AB - In machine learning, positive-unlabelled PU learning is a special case within semi-supervised learning. In positive-unlabelled learning, the training set contains some positive examples and a set of unlabelled examples from both the positive and negative classes. Positive-unlabelled learning has gained attention in many domains, especially in time-series data, in which the obtainment of labelled data is challenging. Examples which originate from the negative class are especially difficult to acquire. Self-learning is a semi-supervised method capable of PU learning in time-series data. In the self-learning approach, observations are individually added from the unlabelled data into the positive class until a stopping criterion is reached. The model is retrained after each addition with the existent labels. The main problem in self-learning is to know when to stop the learning. There are multiple, different stopping criteria in the literature, but they tend to be inaccurate or challenging to apply. This publication proposes a novel stopping criterion, which is called Peak evaluation using perceptually important points, to address this problem for time-series data. Peak evaluation using perceptually important points is exceptional, as it does not have tunable hyperparameters, which makes it easily applicable to an unsupervised setting. Simultaneously, it is flexible as it does not make any assumptions on the balance of the dataset between the positive and the negative class.

UR - http://www.scopus.com/inward/record.url?scp=85074983963&partnerID=8YFLogxK

U2 - 10.1109/JAS.2019.1911777

DO - 10.1109/JAS.2019.1911777

M3 - Article

AN - SCOPUS:85074983963

VL - 6

SP - 1318

EP - 1331

JO - IEEE/CAA Journal of Automatica Sinica

JF - IEEE/CAA Journal of Automatica Sinica

SN - 2329-9266

IS - 6

M1 - 8894745

ER -