Sentiment Analysis of Portuguese Comments from Foursquare
| Autor: | Bruno Á. Souza, Thais G. Almeida, Alice A. F. Menezes, Carlos M. S. Figueiredo, Eduardo F. Nakamura |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
Social network
Language complexity business.industry Computer science Principle of maximum entropy Supervised learning Sentiment analysis 020207 software engineering 02 engineering and technology computer.software_genre Machine learning language.human_language Support vector machine 0202 electrical engineering electronic engineering information engineering language 020201 artificial intelligence & image processing Social media Artificial intelligence Portuguese business computer Natural language processing |
| Zdroj: | WebMedia |
| DOI: | 10.1145/2976796.2988180 |
| Popis: | Sentiment Analysis is an emerging research area applied to social media to find useful information from specific topics, such as service or product quality, or even general contexts as marketing, politics and economy. Although there are numerous studies using Sentiment Analysis, few of them address the Portuguese language, because of the language complexity. This study aims to evaluate the performance of three methods of machine learning (Support Vector Machine, Multinomial Naive Bayes, and Maximum Entropy) to detect polarities in two different scenarios. The first scenario is characterized as a three-class problem (positive, negative, and neutral). The second scenario consists of only two classes (positive and negative). Our dataset consists of comments from the Foursquare social network. The results show that Support Vector Machine presents an F-score at least 1.9% higher than the others for the three-class scenario, while Multinomial Naive Bayes presents an F-score at least 2.6% higher than the others for two-class scenario. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|