Abstract

Dimension reduction techniques form the core of predictive analytics systems and they help us create a new feature space that is more helpful in predicting response variables. But these techniques do not necessarily guarantee a better predictive capability as most of them are unsupervised, especially in regression learning. In regression analysis literature, supervised dimension reduction techniques have not been explored much and in this work we provide a solution to this through probabilistic topic models. In this work, we have shown that the double mixture structure of Latent Dirichlet Allocation topic model helps us 1) to visualize feature patterns, and 2) to project features onto a topic simplex that is more predictive of responses, when compared to popular techniques like PCA and KernelPCA. Until now, topic models have not been explored in a supervised context of video analysis and in this work we introduce a Regularized supervised topic model (RSLDA) that models video and audio features and has outperformed supervised dimension reduction techniques like SPCA and Correlation based feature selection algorithms. All the models discussed in this work have been evaluated to predict continuous human emotions from video data.

Original languageEnglish (US)
Title of host publicationProceedings - 2015 IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages739-744
Number of pages6
ISBN (Electronic)9781509002870
DOIs
StatePublished - Mar 2 2016
EventIEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015 - Miami, United States
Duration: Dec 9 2015Dec 11 2015

Other

OtherIEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015
CountryUnited States
CityMiami
Period12/9/1512/11/15

Fingerprint

Regression analysis
Feature extraction
Statistical Models
Predictive analytics

Keywords

  • Dimension reduction
  • Regression analysis
  • Topic models

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications

Cite this

Lade, P., Demakethepalli Venkateswara, H., & Panchanathan, S. (2016). Regularized supervised topic model for continuous emotion analysis. In Proceedings - 2015 IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015 (pp. 739-744). [7424409] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICMLA.2015.219

Regularized supervised topic model for continuous emotion analysis. / Lade, Prasanth; Demakethepalli Venkateswara, Hemanth; Panchanathan, Sethuraman.

Proceedings - 2015 IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015. Institute of Electrical and Electronics Engineers Inc., 2016. p. 739-744 7424409.

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

Lade, P, Demakethepalli Venkateswara, H & Panchanathan, S 2016, Regularized supervised topic model for continuous emotion analysis. in Proceedings - 2015 IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015., 7424409, Institute of Electrical and Electronics Engineers Inc., pp. 739-744, IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015, Miami, United States, 12/9/15. https://doi.org/10.1109/ICMLA.2015.219
Lade P, Demakethepalli Venkateswara H, Panchanathan S. Regularized supervised topic model for continuous emotion analysis. In Proceedings - 2015 IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015. Institute of Electrical and Electronics Engineers Inc. 2016. p. 739-744. 7424409 https://doi.org/10.1109/ICMLA.2015.219
Lade, Prasanth ; Demakethepalli Venkateswara, Hemanth ; Panchanathan, Sethuraman. / Regularized supervised topic model for continuous emotion analysis. Proceedings - 2015 IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015. Institute of Electrical and Electronics Engineers Inc., 2016. pp. 739-744
@inproceedings{33b097e6d2e14a0f8eed6a7c160567e8,
title = "Regularized supervised topic model for continuous emotion analysis",
abstract = "Dimension reduction techniques form the core of predictive analytics systems and they help us create a new feature space that is more helpful in predicting response variables. But these techniques do not necessarily guarantee a better predictive capability as most of them are unsupervised, especially in regression learning. In regression analysis literature, supervised dimension reduction techniques have not been explored much and in this work we provide a solution to this through probabilistic topic models. In this work, we have shown that the double mixture structure of Latent Dirichlet Allocation topic model helps us 1) to visualize feature patterns, and 2) to project features onto a topic simplex that is more predictive of responses, when compared to popular techniques like PCA and KernelPCA. Until now, topic models have not been explored in a supervised context of video analysis and in this work we introduce a Regularized supervised topic model (RSLDA) that models video and audio features and has outperformed supervised dimension reduction techniques like SPCA and Correlation based feature selection algorithms. All the models discussed in this work have been evaluated to predict continuous human emotions from video data.",
keywords = "Dimension reduction, Regression analysis, Topic models",
author = "Prasanth Lade and {Demakethepalli Venkateswara}, Hemanth and Sethuraman Panchanathan",
year = "2016",
month = "3",
day = "2",
doi = "10.1109/ICMLA.2015.219",
language = "English (US)",
pages = "739--744",
booktitle = "Proceedings - 2015 IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Regularized supervised topic model for continuous emotion analysis

AU - Lade, Prasanth

AU - Demakethepalli Venkateswara, Hemanth

AU - Panchanathan, Sethuraman

PY - 2016/3/2

Y1 - 2016/3/2

N2 - Dimension reduction techniques form the core of predictive analytics systems and they help us create a new feature space that is more helpful in predicting response variables. But these techniques do not necessarily guarantee a better predictive capability as most of them are unsupervised, especially in regression learning. In regression analysis literature, supervised dimension reduction techniques have not been explored much and in this work we provide a solution to this through probabilistic topic models. In this work, we have shown that the double mixture structure of Latent Dirichlet Allocation topic model helps us 1) to visualize feature patterns, and 2) to project features onto a topic simplex that is more predictive of responses, when compared to popular techniques like PCA and KernelPCA. Until now, topic models have not been explored in a supervised context of video analysis and in this work we introduce a Regularized supervised topic model (RSLDA) that models video and audio features and has outperformed supervised dimension reduction techniques like SPCA and Correlation based feature selection algorithms. All the models discussed in this work have been evaluated to predict continuous human emotions from video data.

AB - Dimension reduction techniques form the core of predictive analytics systems and they help us create a new feature space that is more helpful in predicting response variables. But these techniques do not necessarily guarantee a better predictive capability as most of them are unsupervised, especially in regression learning. In regression analysis literature, supervised dimension reduction techniques have not been explored much and in this work we provide a solution to this through probabilistic topic models. In this work, we have shown that the double mixture structure of Latent Dirichlet Allocation topic model helps us 1) to visualize feature patterns, and 2) to project features onto a topic simplex that is more predictive of responses, when compared to popular techniques like PCA and KernelPCA. Until now, topic models have not been explored in a supervised context of video analysis and in this work we introduce a Regularized supervised topic model (RSLDA) that models video and audio features and has outperformed supervised dimension reduction techniques like SPCA and Correlation based feature selection algorithms. All the models discussed in this work have been evaluated to predict continuous human emotions from video data.

KW - Dimension reduction

KW - Regression analysis

KW - Topic models

UR - http://www.scopus.com/inward/record.url?scp=84969590914&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84969590914&partnerID=8YFLogxK

U2 - 10.1109/ICMLA.2015.219

DO - 10.1109/ICMLA.2015.219

M3 - Conference contribution

SP - 739

EP - 744

BT - Proceedings - 2015 IEEE 14th International Conference on Machine Learning and Applications, ICMLA 2015

PB - Institute of Electrical and Electronics Engineers Inc.

ER -