SINAI/MCE-Corpus

Dataset Description

Paper: Sentiment polarity detection in Spanish reviews combining supervised and unsupervised approaches

Point of Contact: jmperea@ujaen.es, emcamara@ujaen.es

MuchoCine corpus in English (MCE) is the translated version of the MuchoCine corpus (Spanish Movies Reviews). The MuchoCine corpus was developed by the researcher Fermín Cruz Mata and presented in 2008 at number 41 of the journal Natural Language Processing in the paper titled Document Classification based on Opinion: experiments with a corpus of Spanish cinema reviews.

This paper Sentiment polarity detection in Spanish reviews combining supervised and unsupervised approaches checks the validity of a methodology for polarity classification in Spanish which consists of combining three classifiers, two of them supervised (on texts in English and another language) and an unsupervised classifier using some English language resource for sentiment analysis. This methodology was previously proposed for opinions in Arabic in the paper Improving Polarity Classification of Bilingual Parallel Corpora combining Machine Learning and Semantic Orientation approaches (in press).

The polarity of the documents of the corpus are measured on a scale of 1 to 5, with 1 being very bad and 5 very good.

Licensing Information

MCE is released under the Apache-2.0 License.

Citation Information

@article{MARTINVALDIVIA20133934,
title = {Sentiment polarity detection in Spanish reviews combining supervised and unsupervised approaches},
journal = {Expert Systems with Applications},
volume = {40},
number = {10},
pages = {3934-3942},
year = {2013},
issn = {0957-4174},
doi = {https://doi.org/10.1016/j.eswa.2012.12.084},
url = {https://www.sciencedirect.com/science/article/pii/S0957417412013267},
author = {María-Teresa Martín-Valdivia and Eugenio Martínez-Cámara and Jose-M. Perea-Ortega and L. Alfonso Ureña-López},
keywords = {Sentiment polarity detection, Multilingual opinion mining, Spanish review corpus, SentiWordNet, Metaclassifiers, Stacking algorithm, Voting system},
abstract = {Sentiment polarity detection is one of the most popular tasks related to Opinion Mining. Many papers have been presented describing one of the two main approaches used to solve this problem. On the one hand, a supervised methodology uses machine learning algorithms when training data exist. On the other hand, an unsupervised method based on a semantic orientation is applied when linguistic resources are available. However, few studies combine the two approaches. In this paper we propose the use of meta-classifiers that combine supervised and unsupervised learning in order to develop a polarity classification system. We have used a Spanish corpus of film reviews along with its parallel corpus translated into English. Firstly, we generate two individual models using these two corpora and applying machine learning algorithms. Secondly, we integrate SentiWordNet into the English corpus, generating a new unsupervised model. Finally, the three systems are combined using a meta-classifier that allows us to apply several combination algorithms such as voting system or stacking. The results obtained outperform those obtained using the systems individually and show that this approach could be considered a good strategy for polarity classification when we work with parallel corpora.}
}