Improved method of character traits based on Wechat dataset

In recent years the communication tool Wechat is used widely, however few attention to the academic research. Moreover, there are very few studies about identifying character traits from its data. Previously a method based on data features to decide character traits was proposed in our paper. Here an improve method of the feature set is given through refining new features, and building the feature subset employing feature extracting algorithm. Then it compares the contributions of the original feature set with the feature subset through several classification algorithms. Experimental results show that the classification based on data features can efficiently identify character traits.