Stop words in review summarization using TextRank

This paper presents a comparison of automatic review summarization with and without stop words. An extractive, unsupervised graph-based ranking model TextRank is employed to highlight the differences between both approaches. Experimental results on 50 sample reviews have shown that the usage of stop words removal can be impactful in determining the result of review summarization, which suggests that depending on the user requirements, it should be considered whether stop words removal needs to be performed or not.