Sentiment analysis of scientific citation
Citation sentiment analysis is an application of sentiment analysis in citation content analysis, which aims to determine the sentiment polarity that the citation context carries towards the cited paper. In citation sentiment analysis, the opinion target is the cited paper, and the categories of sentiment polarity could be either positive, negative or neutral. A scientific citation is a reference made to a scientific publication which includes a book or article or technical papers, etc. These are related to the topic of discussion in the cited text.
The extraction of the sentiments can be different from traditional opinion mining from reviews/tweets. The citations are made in a formal language and the use of strong sentiment is generally avoided. Hence, the usual negative words like 'bad', 'dislike', 'disagree', etc are not used. Instead, the same sentiment is expressed in a subtle way such as 'outerform', 'rather unexplored',etc. Thus, an algorithm to classify the sentiment expressed in the citation can be a challenging task.
Given the difficulties with the expression of sentiment by scientific citations, This project attempts to find out if an automatic analysis by an algorithm will be possible.
Labelled dataset is obtained from : https://cl.awaisathar.com/citation-sentiment-corpus/