Evaluating the Impact of a Two-Stage Multivariate Data Cleansing Approach to Improve to the Performance of Machine Learning Classifiers: A Case Study in Human Activity Recognition

Dionicio Neira-Rodado; CD Nugent; I Cleland; Javier  Velasquez; Amelec Viloria

doi:10.3390/s20071858

Evaluating the Impact of a Two-Stage Multivariate Data Cleansing Approach to Improve to the Performance of Machine Learning Classifiers: A Case Study in Human Activity Recognition

Dionicio Neira-Rodado, CD Nugent, I Cleland, Javier Velasquez, Amelec Viloria

Research output: Contribution to journal › Article › peer-review

8 Citations (Scopus)

118 Downloads (Pure)

Abstract

Human activity recognition (HAR) is a popular field of study. The outcomes of the projects in this area have the potential to impact on the quality of life of people with conditions such as dementia. HAR is focused primarily on applying machine learning classifiers on data from low level sensors such as accelerometers. The performance of these classifiers can be improved through an adequate training process. In order to improve the training process, multivariate outlier detection was used in order to improve the quality of data in the training set and, subsequently, performance of the classifier. The impact of the technique was evaluated with KNN and random forest (RF) classifiers. In the case of KNN, the performance of the classifier was improved from 55.9% to 63.59%.

Original language	English
Article number	1858
Pages (from-to)	1-23
Number of pages	23
Journal	Sensors
Volume	20
Issue number	7
Early online date	27 Mar 2020
DOIs	https://doi.org/10.3390/s20071858
Publication status	Published online - 27 Mar 2020

Bibliographical note

Funding Information:
European Union's Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No. 734355.

Publisher Copyright:
© 2020 by the authors. Licensee MDPI, Basel, Switzerland.

Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.

Keywords

Dataset quality
HAR
Machine learning
Multivariate analysis

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.3390/s20071858Licence: CC BY

Final published versionFinal published version, 1.29 MBLicence: CC BY

Cite this

@article{e812b86e12d84d4889d3d6b5b3d7e714,

title = "Evaluating the Impact of a Two-Stage Multivariate Data Cleansing Approach to Improve to the Performance of Machine Learning Classifiers: A Case Study in Human Activity Recognition",

abstract = "Human activity recognition (HAR) is a popular field of study. The outcomes of the projects in this area have the potential to impact on the quality of life of people with conditions such as dementia. HAR is focused primarily on applying machine learning classifiers on data from low level sensors such as accelerometers. The performance of these classifiers can be improved through an adequate training process. In order to improve the training process, multivariate outlier detection was used in order to improve the quality of data in the training set and, subsequently, performance of the classifier. The impact of the technique was evaluated with KNN and random forest (RF) classifiers. In the case of KNN, the performance of the classifier was improved from 55.9% to 63.59%.",

keywords = "Dataset quality, HAR, Machine learning, Multivariate analysis",

author = "Dionicio Neira-Rodado and CD Nugent and I Cleland and Javier Velasquez and Amelec Viloria",

note = "Funding Information: European Union's Horizon 2020 research and innovation program under the Marie Sk{\l}odowska-Curie grant agreement No. 734355. Publisher Copyright: {\textcopyright} 2020 by the authors. Licensee MDPI, Basel, Switzerland. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.",

year = "2020",

month = mar,

day = "27",

doi = "10.3390/s20071858",

language = "English",

volume = "20",

pages = "1--23",

journal = "Sensors",

issn = "1424-8220",

publisher = "MDPI",

number = "7",

}

TY - JOUR

T1 - Evaluating the Impact of a Two-Stage Multivariate Data Cleansing Approach to Improve to the Performance of Machine Learning Classifiers: A Case Study in Human Activity Recognition

AU - Neira-Rodado, Dionicio

AU - Nugent, CD

AU - Cleland, I

AU - Velasquez, Javier

AU - Viloria, Amelec

N1 - Funding Information: European Union's Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No. 734355. Publisher Copyright: © 2020 by the authors. Licensee MDPI, Basel, Switzerland. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.

PY - 2020/3/27

Y1 - 2020/3/27

N2 - Human activity recognition (HAR) is a popular field of study. The outcomes of the projects in this area have the potential to impact on the quality of life of people with conditions such as dementia. HAR is focused primarily on applying machine learning classifiers on data from low level sensors such as accelerometers. The performance of these classifiers can be improved through an adequate training process. In order to improve the training process, multivariate outlier detection was used in order to improve the quality of data in the training set and, subsequently, performance of the classifier. The impact of the technique was evaluated with KNN and random forest (RF) classifiers. In the case of KNN, the performance of the classifier was improved from 55.9% to 63.59%.

AB - Human activity recognition (HAR) is a popular field of study. The outcomes of the projects in this area have the potential to impact on the quality of life of people with conditions such as dementia. HAR is focused primarily on applying machine learning classifiers on data from low level sensors such as accelerometers. The performance of these classifiers can be improved through an adequate training process. In order to improve the training process, multivariate outlier detection was used in order to improve the quality of data in the training set and, subsequently, performance of the classifier. The impact of the technique was evaluated with KNN and random forest (RF) classifiers. In the case of KNN, the performance of the classifier was improved from 55.9% to 63.59%.

KW - Dataset quality

KW - HAR

KW - Machine learning

KW - Multivariate analysis

UR - https://pure.ulster.ac.uk/en/publications/evaluating-the-impact-of-a-two-stage-multivariate-data-cleansing-

UR - http://www.scopus.com/inward/record.url?scp=85082791314&partnerID=8YFLogxK

U2 - 10.3390/s20071858

DO - 10.3390/s20071858

M3 - Article

SN - 1424-8220

VL - 20

SP - 1

EP - 23

JO - Sensors

JF - Sensors

IS - 7

M1 - 1858

ER -

Evaluating the Impact of a Two-Stage Multivariate Data Cleansing Approach to Improve to the Performance of Machine Learning Classifiers: A Case Study in Human Activity Recognition

Abstract

Bibliographical note

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this