Analyzing Emotion in Blog and News at Word and Sentence Level.
Dipankar Das, Sivaji Bandyopadhyay · Indian International Conference on Artificial Intelligence · 2009
Emotion is crucial to identify as it is not open to any objective observation or verification. In this paper, emotion analysis on blog texts has been carried out for a less privileged language, Bengali and the same system has been applied on the English SemEval 2007 affect sensing corpus containing only news headlines. A set of six emotion tags, namely, happy, sad, anger, fear, disgust and surprise, have been selected towards this emotion detection task for reliable and semi-automatic annotation of blog and news data. Conditional Random Field (CRF) based classifier has been applied for recognizing six basic emotion tags for different words of a sentence. The classifier accuracy has been improved by arranging an equal distribution of emotional tags and nonemotional tag. A score based technique has been adopted to calculate and assign tag weights to each of the six emotion tags. A sense based scoring strategy has been applied to identify sentence level emotion scores for the six emotion tags based on the acquired word level emotion tags. Sentence level emotion tagging has been carried out based on the maximum obtained sentence level emotion scores. Evaluation has been conducted for each emotion class separately on 200 test sentences from each of the Bengali blog and English news data. The system has resulted accuracies of 64.28% and 65.80% for happy, 66.42% and 60.42% for sad, 60.28% and 59.28% for anger, 72.18% and 63.60% for disgust, 67.14% and 67.54% for fear and 66.45% and 61.14% for surprise emotion classes on blog and news test data respectively.