coarse-grained sentiment analysis.csv

This study selects the customer service conversation dataset from Jing Dong (JD)[1] to evaluate our proposed model. In the dataset, the number 0 represents the statement issued by the customer, the number 1 represents the statement issued by customer service and chatbots, and the "session-id" represents multi-turn sessions between the same customer service agents and customers. This paper selected about 10,000 dialogue texts for training, validation, and testing, including 5,068 sentences issued by customers and 4,942 sentences issued by customer service and chatbots.<br> [1] https://www.jd.com