Premium
Identifying factors impacting entity sentiment analysis: A case study of sentiment analysis in the context of news reports
Author(s) -
Luo Manman,
Mu Xiangming
Publication year - 2020
Publication title -
proceedings of the association for information science and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.193
H-Index - 14
ISSN - 2373-9231
DOI - 10.1002/pra2.382
Subject(s) - sentiment analysis , computer science , schema (genetic algorithms) , exploratory analysis , coding (social sciences) , information retrieval , context (archaeology) , set (abstract data type) , natural language processing , data science , paleontology , statistics , mathematics , biology , programming language
Entity sentiment analysis aims to analyze sentiment for a specific entity. As entity sentiment analysis requires the understanding of entities and associated entities, this paper presents preliminary findings from an exploratory analysis of factors affecting entity sentiment analysis in the news report context using a data set of 8,754 paragraphs. An entity coding schema was developed based on grounded theory. The coding result shows that 29% of paragraphs express sentiment directly, and 24% express indirectly. Negative biased entities (towards target entity) may significantly affect the accuracy of entity sentiment analysis. The findings may shed light on future exploration in improving the accuracy of entity sentiment analysis.