Paper Title
Emotion Generation and Summarization Form Affective Text

Abstract
Mining social emotion from text deals with new aspect for categorizing the document based on the emotions such as victory, love, anger etc. In order to predict the emotion contained in content a joint emotion-topic model is proposed by enhancing Latent Dirichlet Allocation with an additional layer for emotion modeling. Using this it first generates a latent topic from emotions, followed by generating perceptual terms from each topic. First it generates an emotion from a document-specific emotional distribution, and then it generates a latent topic from a multinominal distribution conditioned on emotions. The model which we proposed will utilize the complementary advantages of both emotion-term model and topic model and also it include more websites for creating a large vocabulary. Emotion-topic model allows associating the terms i.e. words and emotions via topics which is more flexible. Also it has better modeling capability. For each emotion, it generates a meaningful latent topic and also based on emotions, songs recommendation will be available for user. So that user can upload and enjoy their own choice of song. Keywords— Affective Text Mining, Emotional-Term Model.