Enrichment of dictionaries to improve the automatic classification of feelings in postings related to the use of systems

Abstract

This work proposes an investigation to improve the efficiency of a lexical-based classifier, the SentiStrength, for automatic sentiment detection in postings related to the use of systems. To achieve this goal, the TF-IDF metric was used to select words that are related to the domain of the posts, which will enrich the dictionary used by the tool to generate the polarity of the posts. The efficiency of a dictionarie enriched with words in their root form and a dictionarie enriched with lematized words will also be investigated. The research was conducted with 2108 sentences extracted from the reviews section of the Play Store on urban mobility applications, such as Waze, Google Maps and GPS Brazil. One of the results obtained was a 7.3 % increase in the accuracy of the classifier when using enriched dictionaries.