A book reviews dataset for Aspect-Based Sentiment Analysis
ABSTRACT
Aspect-based sentiment analysis (ABSA) is the task of identifying specific entities mentioned in an opinionated text and the sentiment associated to them. Most of the research carried out on this subject is based on supervised approaches, which means requiring a considerable amount of tagged data for training. However, there are no many of these corpora, as their creation requires a lot of time and resources. For this reason, our main contribution in this paper is the development of a new manually tagged dataset for ABSA in English for the context of book reviews. This dataset identifies the different aspects and polarities in each sentence of a review. Moreover, we defined several categories related to books, which are also identified and every aspect is classified into one of them. This new corpus contains a total number of 2977 sentences, belonging to 300 reviews of 40 different books and the total number of aspects identified is 3833, including both explicit and implicit aspects.
Files in this item
- Name:
- 2017_alvarez_book_reviews.pdf
- Size:
- 5.028Mb
- Format:
- Description:
- Embargo indefinido por copyright