Abstract: Sentiment analysis is used to identify and extract subjective information (verb, noun, adjectives) in text sentences. The dataset used for emotion identification is collected from twitter. User posting contents in twitter is increasing every day. Emotion identification is one of the concept for understanding users behaviour, action and is important to all phases of our life. The emotions are automatically labelled according to the emotion hashtag. Each information recovered from twitter has separate hashtag. Naive Bayes (NB) machine learning algorithm is used for emotion identification. To find valuable features for emotion identification like n-grams, LIWC Dictionary, part-of-speech (POS). Finally the extracted Features is fed into a Naive Bayes (NB) classifier to attain classification accuracy on emotion identification in Twitter dataset. Na´ve Bayes classification is a simple probabilistic model that works well on text classification. By using Twitter content, it classifies the different emotions like joy, sadness, anger, love, fear, thankfulness, and surprise for the twitter content.

Keywords: Na´ve Bayes(NB) classifier,N-Gram, LIWC Dictionary, Part-of-speech.