Predicting personality traits of Chinese users based on Facebook wall posts

摘要

Automatically recognizing personality based on historical action logs in online social networks is a promising method to infer a person's behaviors, and it has received a lot of attention lately as it might lead to the construction of a better personal recommendation system. However, very few previous works in the literature put their focus on predicting personality from Chinese texts. As Chinese texts are much more difficult to delimit than English texts, it poses more challenges in recognizing personality from Chinese texts. In this paper, we attempt to classify the personality traits from Chinese texts. We collected a dataset with posts and personality scores of 222 Facebook users who use Chinese as their main written language. Then, we used Jieba, a Chinese text segmentation tool, as the tokenizer for the task of text segmentation, and the Support Vector Machine (SVM) as the learning algorithm for personality classification. Our experimental results show that the performance in precision and recall can be significantly improved with the help of text segmentation. Moreover, exploiting side information, such as the number of friends, could improve the performance further. One interesting finding from our experiments is that extraverts seem to write more sentences and use more common words than introverts. This indicates that extraverts are more willing to share their mood and life with others than introverts.