This paper fills a gap in aspect-based sentiment analysis and aims to presenta new method for preparing and analysing texts concerning opinion andgenerating user-friendly descriptive reports in natural language. We present acomprehensive set of techniques derived from Rhetorical Structure Theory andsentiment analysis to extract aspects from textual opinions and then build anabstractive summary of a set of opinions. Moreover, we propose aspect-aspectgraphs to evaluate the importance of aspects and to filter out unimportant onesfrom the summary. Additionally, the paper presents a prototype solution of dataflow with interesting and valuable results. The proposed method's resultsproved the high accuracy of aspect detection when applied to the gold standarddataset.
展开▼