Chinese_conversation_sentiment

A Chinese sentiment dataset may be useful for sentiment analysis.

sentiment_XS_test.txt contains 11577 instances labeled manually (XS_test referred in the paper). sentiment_XS_30k.txt contains almost 30k instances labeled automatically (XS_30k referred in the paper).

All data are from human-computer conversation logs and are segmented by Jieba segmentation tool.

If you use this dataset, please cite paper: Sentiment Classification with Convolutional Neural Networks: an Experimental Study on a Large-scale Chinese Conversation Corpus, in the 12th International Conference on Computational Intelligence and Security (CIS2016)

Contact me: [email protected]