Evaluation of sentiment polarity prediction using a dimensional and a categorical approach

I. Muresan; A. Stan; M. Giurgiu; R. Potolea

doi:10.1109/SpeD.2013.6682645

Evaluation of sentiment polarity prediction using a dimensional and a categorical approach

I. Muresan, A. Stan, M. Giurgiu, R. Potolea

Institute of Language, Cognition and Computation

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

In this paper we evaluate two approaches for predicting the sentiment polarity of an utterance. The first method is based on a 3-dimensional model which takes into account text expressiveness in terms of valence, arousal and dominance. The second one determines the word's semantic orientation according to Chi-square and Relevance factor statistic metrics. We describe the general flow of the methods and their extracted features, as well as their predictability potential using different machine learning algorithms, Naïve Bayes, SVM and C4.5. The evaluation is performed on four emotional datasets: Semeval 2007 “Affective Text”, ISEAR (International Survey on Emotional Antecedents and Reactions), children's fairy-tales and a movie review dataset. The results show a high correlation of the prediction performance with the database content, as well as to the average number of words within the classified text instances.

Original language	English
Title of host publication	Speech Technology and Human - Computer Dialogue (SpeD), 2013 7th Conference on
Pages	1-6
Number of pages	6
DOIs	https://doi.org/10.1109/SpeD.2013.6682645
Publication status	Published - 1 Oct 2013

Keywords / Materials (for Non-textual outputs)

Bayes methods
emotion recognition
feature extraction
learning (artificial intelligence)
pattern classification
statistical analysis
text analysis
C4.5
ISEAR
International Survey on Emotional Antecedents and Reactions
SVM
Semeval Affective Text
arousal
categorical approach
chi-square statistic metrics
children fairy-tales
database content
dimensional approach
dominance
emotional dataset
machine learning algorithm
movie review dataset
naive Bayes
predictability potential
relevance factor statistic metrics
sentiment polarity prediction
text expressiveness
text instance classification
utterance sentiment polarity
valence
word semantic orientation
Databases
Feature extraction
Measurement
Motion pictures
Numerical models
Semantics
Support vector machines
VAD model
sentiment polarity
statistic metrics

Access to Document

10.1109/SpeD.2013.6682645

http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6682645

HELP4MOOD:A computational distributed system to support the treatment of patients with major depression
Matheson, C. (Principal Investigator) & Wolters, M. (Co-Investigator (External))
EU government bodies
1/01/11 → 30/06/14
Project: Research

Cite this

@inproceedings{28d25add392742f0979660d1c481fd57,

title = "Evaluation of sentiment polarity prediction using a dimensional and a categorical approach",

abstract = "In this paper we evaluate two approaches for predicting the sentiment polarity of an utterance. The first method is based on a 3-dimensional model which takes into account text expressiveness in terms of valence, arousal and dominance. The second one determines the word's semantic orientation according to Chi-square and Relevance factor statistic metrics. We describe the general flow of the methods and their extracted features, as well as their predictability potential using different machine learning algorithms, Na{\"i}ve Bayes, SVM and C4.5. The evaluation is performed on four emotional datasets: Semeval 2007 “Affective Text”, ISEAR (International Survey on Emotional Antecedents and Reactions), children's fairy-tales and a movie review dataset. The results show a high correlation of the prediction performance with the database content, as well as to the average number of words within the classified text instances.",

keywords = "Bayes methods, emotion recognition, feature extraction, learning (artificial intelligence), pattern classification, statistical analysis, text analysis, C4.5, ISEAR, International Survey on Emotional Antecedents and Reactions, SVM, Semeval Affective Text, arousal, categorical approach, chi-square statistic metrics, children fairy-tales, database content, dimensional approach, dominance, emotional dataset, machine learning algorithm, movie review dataset, naive Bayes, predictability potential, relevance factor statistic metrics, sentiment polarity prediction, text expressiveness, text instance classification, utterance sentiment polarity, valence, word semantic orientation, Databases, Feature extraction, Measurement, Motion pictures, Numerical models, Semantics, Support vector machines, VAD model, sentiment polarity, statistic metrics",

author = "I. Muresan and A. Stan and M. Giurgiu and R. Potolea",

year = "2013",

month = oct,

day = "1",

doi = "10.1109/SpeD.2013.6682645",

language = "English",

pages = "1--6",

booktitle = "Speech Technology and Human - Computer Dialogue (SpeD), 2013 7th Conference on",

}

TY - GEN

T1 - Evaluation of sentiment polarity prediction using a dimensional and a categorical approach

AU - Muresan, I.

AU - Stan, A.

AU - Giurgiu, M.

AU - Potolea, R.

PY - 2013/10/1

Y1 - 2013/10/1

N2 - In this paper we evaluate two approaches for predicting the sentiment polarity of an utterance. The first method is based on a 3-dimensional model which takes into account text expressiveness in terms of valence, arousal and dominance. The second one determines the word's semantic orientation according to Chi-square and Relevance factor statistic metrics. We describe the general flow of the methods and their extracted features, as well as their predictability potential using different machine learning algorithms, Naïve Bayes, SVM and C4.5. The evaluation is performed on four emotional datasets: Semeval 2007 “Affective Text”, ISEAR (International Survey on Emotional Antecedents and Reactions), children's fairy-tales and a movie review dataset. The results show a high correlation of the prediction performance with the database content, as well as to the average number of words within the classified text instances.

AB - In this paper we evaluate two approaches for predicting the sentiment polarity of an utterance. The first method is based on a 3-dimensional model which takes into account text expressiveness in terms of valence, arousal and dominance. The second one determines the word's semantic orientation according to Chi-square and Relevance factor statistic metrics. We describe the general flow of the methods and their extracted features, as well as their predictability potential using different machine learning algorithms, Naïve Bayes, SVM and C4.5. The evaluation is performed on four emotional datasets: Semeval 2007 “Affective Text”, ISEAR (International Survey on Emotional Antecedents and Reactions), children's fairy-tales and a movie review dataset. The results show a high correlation of the prediction performance with the database content, as well as to the average number of words within the classified text instances.

KW - Bayes methods

KW - emotion recognition

KW - feature extraction

KW - learning (artificial intelligence)

KW - pattern classification

KW - statistical analysis

KW - text analysis

KW - C4.5

KW - ISEAR

KW - International Survey on Emotional Antecedents and Reactions

KW - SVM

KW - Semeval Affective Text

KW - arousal

KW - categorical approach

KW - chi-square statistic metrics

KW - children fairy-tales

KW - database content

KW - dimensional approach

KW - dominance

KW - emotional dataset

KW - machine learning algorithm

KW - movie review dataset

KW - naive Bayes

KW - predictability potential

KW - relevance factor statistic metrics

KW - sentiment polarity prediction

KW - text expressiveness

KW - text instance classification

KW - utterance sentiment polarity

KW - valence

KW - word semantic orientation

KW - Databases

KW - Feature extraction

KW - Measurement

KW - Motion pictures

KW - Numerical models

KW - Semantics

KW - Support vector machines

KW - VAD model

KW - sentiment polarity

KW - statistic metrics

U2 - 10.1109/SpeD.2013.6682645

DO - 10.1109/SpeD.2013.6682645

M3 - Conference contribution

SP - 1

EP - 6

BT - Speech Technology and Human - Computer Dialogue (SpeD), 2013 7th Conference on

ER -

Evaluation of sentiment polarity prediction using a dimensional and a categorical approach

Abstract

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Projects

HELP4MOOD:A computational distributed system to support the treatment of patients with major depression

Cite this