Sentiment analysis of text messages using ensemble of classifiers

• Grant number: 14/01237-7
• Support type: Regular Research Grants
• Duration: May 01, 2014 - April 30, 2016
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal researcher: Tiago Agostinho de Almeida
• Grantee: Tiago Agostinho de Almeida
• Home Institution: Centro de Ciências e Tecnologias para a Sustentabilidade (CCTS). Universidade Federal de São Carlos (UFSCAR). Sorocaba , SP, Brazil

Abstract

The recent growth of social networks and digital inclusion have allowed users to express their opinions through the Internet. This fact has changed the way companies provide customer services and how costumers express their opinions.Nowaday, reports indicate that social networks are one of the most widely used digital media. Some studies show that the time spent in such environments is usually shared with relevant tasks, such as work and study. In this way, due to lack of time, text messages are often written with rife of idioms, symbols and abbreviations commonly employed to compress the original message in order to type quickly and avoid the size limit imposed by the enviroment. Automatically detect the polarity of messages is still a scientific challenge that has attracted the attention of both the market and academy since its application is wide, such as to estimate the popularity of a candidate within his electorate, to infer whether a product is being well rated by their customers, among many others. Beyond such examples, detect the polarity of messages can also help companies understand the opinion of a large amount of customers regarding marketed products. In addition, it could assist consumers intentioned to buy a product know the consensus opinion of other customers.The need to analyze many different sources with large volumes of opinions establishes a potential way for scientific research. In this scenario, this project proposes a system for detecting messages polarity with ensemble and techiniques for increasing the feature space. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

LOCHTER, JOHANNES V.; ZANETTI, RAFAEL F.; RELLER, DOMINIK; ALMEIDA, TIAGO A. Short text opinion detection using ensemble of classifiers and semantic indexing. EXPERT SYSTEMS WITH APPLICATIONS, v. 62, p. 243-249, NOV 15 2016. Web of Science Citations: 10.

ALMEIDA, TIAGO A.; SILVA, TIAGO P.; SANTOS, IGOR; GOMEZ HIDALGO, JOSE M. Text normalization and semantic indexing to enhance Instant Messaging and SMS spam filtering. KNOWLEDGE-BASED SYSTEMS, v. 108, n. SI, p. 25-32, SEP 15 2016. Web of Science Citations: 15.