Loading...
Please wait, while we are loading the content...
Reader ’ s Emotion Prediction Based on Partitioned Latent Dirichlet Allocation Model
| Content Provider | Semantic Scholar |
|---|---|
| Author | Xu, Ruifeng Zou, Chengtian |
| Copyright Year | 2013 |
| Abstract | Different from traditional emotion analysis which focuses on the identification of emotions from the text, this research aims to predict the reader’s emotion for given text. Regarding reader emotion as the response to the text, emotion prediction may be transferred to a classification problem which classifies the text into the categories causing different emotions. In this study, we propose an emotion prediction approach based on Partitioned Latent Dirichlet Allocation (PLDA) model. Through providing the supervised information to the training process of LDA model, PLDA model associates the words from one certain type of emotion to one certain partition of topics. The outputs of PLDA model are used as the features of a multi-label classifier for predicating the reader’s emotion. Evaluations on a large community emotion corpus show that PLDA model achieves much better performance compared to bag of words model |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://weblidi.info.unlp.edu.ar/WorldComp2013-Mirror/p2013/ICM3532.pdf |
| Alternate Webpage(s) | http://worldcomp-proceedings.com/proc/p2013/ICM3532.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
