User intention understanding from scratch

User intention understanding from text is an important task in NLP. In this paper, we study the problem of phone-changing intention prediction. And we propose a novel feature extraction method, which selects the most representative intention feature, to represent user's intention from text scratch. Then we adopt a supervised learning approach, that is to train SVM classifier, for intention prediction. In addition, we propose a novel phone-changing intention dataset that the text scratches and their corresponding labels are collected from real network environment. The experimental results validate the effectiveness of our proposed approach.