A corpus for aspect-based sentiment analysis in Vietnamese

Abstract

Recently, researchers have shown an increased interest in the aspect-based sentiment analysis problem. The goal is to extract valuable information concerning the aspects mentioned in users comments. This problem can be divided into three sub-tasks: term extraction, aspect detection, and polarity detection. In this paper, we present a new annotated corpus for studies on the two sub-tasks: aspect detection and polarity detection. Our corpus includes 7,828 restaurant reviews at document-level. We also performed a supervised learning method with rich features, achieving the F1-score of 87.13% for the aspect detection and the F1-score of 59.20% for polarity detection. Our corpus is published for research purpose1.