期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2014
卷号：3
期号：2
页码：458-462
出版社：Shri Pannalal Research Institute of Technolgy
摘要：The quick development of net a pair additional and additional documents are assigned by social users with feeling labels like happiness, sadness, and surprise. Such emotions will offer a replacement facet for document categorization, and so facilitate on- line users to pick out connected documents supported their emotional preferences. The quantitative relation with manual feeling labels remains terribly small comparison to the large quantity of web/enterprise documents. The connections between social feelings and emotive terms and supported that predict the social emotion from text content mechanically. A joint feeling- topic model by augmenting Latent Dirichlet Allocation with an extra layer for emotion modeling. It generates a group of latent topics from emotions, followed by generating emotive terms from every topic. Experimental results on a web news assortment show that the projected model will effectively determine substantive latent topics for every feeling. The analysis on feeling prediction any verifies the effectiveness of the projected model.
