The dataset comes from the 2014 NLPCC (Natural Language Processing and Chinese Computing Conference), which is an annual academic conference of the Chinese Information Technology Professional Committee hosted by the China Computer Federation (CCF).

The evaluation data comes from Sina Weibo. For microblogs containing emotions, their emotional classification output can be judged as anger, disgust, fear, happiness, like, sadness, and surprise.

The data format is XML, encoded in Unicode (utf-16), and includes: emotion classification, emotion classification ID, emotion expression identification, and expression ID files.