Introduction
-
Chinese dataset: The pre-processed Chinese datasets include dataset1 and dataset2 are collected from BiliBili websites. We use Baidu natural language API to extract sentiment from review. The meaning of column are listed as follow: user, item, time stamp, confidence, positive, negative. where confidence, positive and negative represent the dimension of sentiment, which have been calculated from review.
-
English dataset: The English datasets include bars and restaurant come from Yelp Challenge. We use Standford NLP to extract user's sentiment.The meaning of column are listed as follow: user, item, time stamp, very positive, positive, neutral, negative, very negative.
Our paper have accepted by International Journal of Data Science and Analytics. You can acquire it at paper address or access this url: https://doi.org/10.1007/s41060-022-00374-5.
Department of Computer Science, Shantou University, China.