Information Technology Journal1812-56381812-5646Asian Network for Scientific Information10.3923/itj.2013.510.513WeijieYang YapengZhu 32013123This paper proposed a method of polarity analysis of stock comments in Chinese based on word clustering, to avoid effect of noisy information without emotional tendency. This method mines semantic association between words to construct the concept dictionary in specific areas dynamically and describes the text feature with the concept constructed. The experimental results show that this method can not only reduces dimensionality of feature space, speed up the process of sentiment classification, but also improve the performance of sentiment classification to a certain extent.]]>Hatzivassiloglou, V. and K.R. McKeown,19971997pp: 174181Yang, W.J. and H.B. Qian,20112011pp: 867870Mo, Q., Y.J. Zhang and H.L. Hu,201147222225Xu, L.H., H.F. Lin and Z.H. Yang,20072196100Hu, H.L. and Q. Mo,20095896899Hu, H.,20092009Li, H. and K. Yamanishi,200339521541Pang, J.F., D.B. Bu and S. Bai,2001182326