Answers are grounded in this paper and cite their sources.
In this paper, we introduce the task of targeted aspect-based sentiment. The goal is to extract fine-grained information with respect to mentioned in user comments. This work extends both aspect-based analysis that assumes a single entity per document and targeted analysis that assumes a single sentiment towards a target entity. In, we identify the sentiment towards each aspect of one or more. As a testbed for this task, we introduce the SentiHood dataset, from a question answering (QA) platform where urban neighbourhoods discussed by users. In this context units of text often mention several of one or more neighbourhoods. This is the first time that a generic media platform in this case a QA platform, is used for fine-grained mining. Text coming from QA platforms is far less constrained compared text from review specific platforms which current datasets are based on. We several strong baselines, relying on logistic regression and-of-the-art recurrent neural networks.
No takes yet. Share an insight, caveat, or question.
Saeidi et al. (2016) studied this question.