期刊名称：International Journal of Electronics, Communication and Soft Computing Science and Engineering
印刷版ISSN：2277-9477
出版年度：2018
卷号：11
语种：English
出版社：IJECSCSE
摘要：Short text is different of traditional documents in its shortness and sparsity. Short texts are prevalent on the web, no matter in traditional websites, e.g., webpage titles, text advertisements and image captions, or in emerging social media, e.g., tweets, status messages, and questions in Q&A websites. This paper focused on five different techniques such as Pre-Training, Extended Naive Bayes, IncreSTS, Text Segmentation, bit term topic model (BTM). But some problems exist in each method so to overcome the problems that are given in analysis and discussion, Logistic Regression method is proposed and the ranking model is proposed to transform the input sentence to the sentiment tree with the highest ranking score.
