Addressing the problems of data-centric physiology-affect relations modeling

Roberto Legaspi, Ken Ichi Fukui, Koichi Moriyama, Satoshi Kurihara, Masayuki Numao, Merlin Suarez

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Data-centric affect modeling may render itself restrictive in practical applications for three reasons, namely, it falls short of feature optimization, infers discrete affect classes, and deals with relatively small to average sized datasets. Though it seems practical to use the feature combinations already associated to commonly investigated sensors, there may be other potentially optimal features that can lead to new relations. Secondly, although it seems more realistic to view affect as continuous, it requires using continuous labels that will increase the difficulty of modeling. Lastly, although a large scale dataset reflects a more precise range of values for any given feature, it severely hinders computational efficiency. We address these problems when inferring physiology-affect relations from datasets that contain 2-3 million feature vectors, each with 49 features and labelled with continuous affect values. We employ automatic feature selection to acquire near optimal feature subsets and a fast approximate kNN algorithm to solve the regression problem and cope with the challenge of a large scale dataset. Our results show that high estimation accuracy may be achieved even when the selected feature subset is only about 7% of the original features. May the results here motivate the HCI community to pursue affect modeling without being deterred by large datasets and further the discussions on acquiring optimal features for accurate continuous affect approximation.

Original languageEnglish
Title of host publicationIUI 2010 - Proceedings of the 14th ACM International Conference on Intelligent User Interfaces
Pages21-30
Number of pages10
DOIs
Publication statusPublished - 2010 Apr 26
Externally publishedYes
Event14th ACM International Conference on Intelligent User Interfaces, IUI 2010 - Hong Kong, China
Duration: 2010 Feb 72010 Feb 10

Other

Other14th ACM International Conference on Intelligent User Interfaces, IUI 2010
CountryChina
CityHong Kong
Period10/2/710/2/10

Fingerprint

Physiology
Human computer interaction
Computational efficiency
Set theory
Feature extraction
Labels
Sensors

Keywords

  • Affective computing
  • Machine learning
  • Pattern recognition

ASJC Scopus subject areas

  • Software
  • Human-Computer Interaction

Cite this

Legaspi, R., Fukui, K. I., Moriyama, K., Kurihara, S., Numao, M., & Suarez, M. (2010). Addressing the problems of data-centric physiology-affect relations modeling. In IUI 2010 - Proceedings of the 14th ACM International Conference on Intelligent User Interfaces (pp. 21-30) https://doi.org/10.1145/1719970.1719974

Addressing the problems of data-centric physiology-affect relations modeling. / Legaspi, Roberto; Fukui, Ken Ichi; Moriyama, Koichi; Kurihara, Satoshi; Numao, Masayuki; Suarez, Merlin.

IUI 2010 - Proceedings of the 14th ACM International Conference on Intelligent User Interfaces. 2010. p. 21-30.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Legaspi, R, Fukui, KI, Moriyama, K, Kurihara, S, Numao, M & Suarez, M 2010, Addressing the problems of data-centric physiology-affect relations modeling. in IUI 2010 - Proceedings of the 14th ACM International Conference on Intelligent User Interfaces. pp. 21-30, 14th ACM International Conference on Intelligent User Interfaces, IUI 2010, Hong Kong, China, 10/2/7. https://doi.org/10.1145/1719970.1719974
Legaspi R, Fukui KI, Moriyama K, Kurihara S, Numao M, Suarez M. Addressing the problems of data-centric physiology-affect relations modeling. In IUI 2010 - Proceedings of the 14th ACM International Conference on Intelligent User Interfaces. 2010. p. 21-30 https://doi.org/10.1145/1719970.1719974
Legaspi, Roberto ; Fukui, Ken Ichi ; Moriyama, Koichi ; Kurihara, Satoshi ; Numao, Masayuki ; Suarez, Merlin. / Addressing the problems of data-centric physiology-affect relations modeling. IUI 2010 - Proceedings of the 14th ACM International Conference on Intelligent User Interfaces. 2010. pp. 21-30
@inproceedings{725a7d3983964cd894e8bf9853143073,
title = "Addressing the problems of data-centric physiology-affect relations modeling",
abstract = "Data-centric affect modeling may render itself restrictive in practical applications for three reasons, namely, it falls short of feature optimization, infers discrete affect classes, and deals with relatively small to average sized datasets. Though it seems practical to use the feature combinations already associated to commonly investigated sensors, there may be other potentially optimal features that can lead to new relations. Secondly, although it seems more realistic to view affect as continuous, it requires using continuous labels that will increase the difficulty of modeling. Lastly, although a large scale dataset reflects a more precise range of values for any given feature, it severely hinders computational efficiency. We address these problems when inferring physiology-affect relations from datasets that contain 2-3 million feature vectors, each with 49 features and labelled with continuous affect values. We employ automatic feature selection to acquire near optimal feature subsets and a fast approximate kNN algorithm to solve the regression problem and cope with the challenge of a large scale dataset. Our results show that high estimation accuracy may be achieved even when the selected feature subset is only about 7{\%} of the original features. May the results here motivate the HCI community to pursue affect modeling without being deterred by large datasets and further the discussions on acquiring optimal features for accurate continuous affect approximation.",
keywords = "Affective computing, Machine learning, Pattern recognition",
author = "Roberto Legaspi and Fukui, {Ken Ichi} and Koichi Moriyama and Satoshi Kurihara and Masayuki Numao and Merlin Suarez",
year = "2010",
month = "4",
day = "26",
doi = "10.1145/1719970.1719974",
language = "English",
isbn = "9781605585154",
pages = "21--30",
booktitle = "IUI 2010 - Proceedings of the 14th ACM International Conference on Intelligent User Interfaces",

}

TY - GEN

T1 - Addressing the problems of data-centric physiology-affect relations modeling

AU - Legaspi, Roberto

AU - Fukui, Ken Ichi

AU - Moriyama, Koichi

AU - Kurihara, Satoshi

AU - Numao, Masayuki

AU - Suarez, Merlin

PY - 2010/4/26

Y1 - 2010/4/26

N2 - Data-centric affect modeling may render itself restrictive in practical applications for three reasons, namely, it falls short of feature optimization, infers discrete affect classes, and deals with relatively small to average sized datasets. Though it seems practical to use the feature combinations already associated to commonly investigated sensors, there may be other potentially optimal features that can lead to new relations. Secondly, although it seems more realistic to view affect as continuous, it requires using continuous labels that will increase the difficulty of modeling. Lastly, although a large scale dataset reflects a more precise range of values for any given feature, it severely hinders computational efficiency. We address these problems when inferring physiology-affect relations from datasets that contain 2-3 million feature vectors, each with 49 features and labelled with continuous affect values. We employ automatic feature selection to acquire near optimal feature subsets and a fast approximate kNN algorithm to solve the regression problem and cope with the challenge of a large scale dataset. Our results show that high estimation accuracy may be achieved even when the selected feature subset is only about 7% of the original features. May the results here motivate the HCI community to pursue affect modeling without being deterred by large datasets and further the discussions on acquiring optimal features for accurate continuous affect approximation.

AB - Data-centric affect modeling may render itself restrictive in practical applications for three reasons, namely, it falls short of feature optimization, infers discrete affect classes, and deals with relatively small to average sized datasets. Though it seems practical to use the feature combinations already associated to commonly investigated sensors, there may be other potentially optimal features that can lead to new relations. Secondly, although it seems more realistic to view affect as continuous, it requires using continuous labels that will increase the difficulty of modeling. Lastly, although a large scale dataset reflects a more precise range of values for any given feature, it severely hinders computational efficiency. We address these problems when inferring physiology-affect relations from datasets that contain 2-3 million feature vectors, each with 49 features and labelled with continuous affect values. We employ automatic feature selection to acquire near optimal feature subsets and a fast approximate kNN algorithm to solve the regression problem and cope with the challenge of a large scale dataset. Our results show that high estimation accuracy may be achieved even when the selected feature subset is only about 7% of the original features. May the results here motivate the HCI community to pursue affect modeling without being deterred by large datasets and further the discussions on acquiring optimal features for accurate continuous affect approximation.

KW - Affective computing

KW - Machine learning

KW - Pattern recognition

UR - http://www.scopus.com/inward/record.url?scp=77951119219&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77951119219&partnerID=8YFLogxK

U2 - 10.1145/1719970.1719974

DO - 10.1145/1719970.1719974

M3 - Conference contribution

AN - SCOPUS:77951119219

SN - 9781605585154

SP - 21

EP - 30

BT - IUI 2010 - Proceedings of the 14th ACM International Conference on Intelligent User Interfaces

ER -