Abstract: The most emerging area in NLP now a days is Sentiment Analysis (SA) which is a cognitive process in which the user’s feelings and emotions are extracted. It has a variety of applications. It can be used to analyze whether the product review is positive or negative, based on tweets how people respond to adds, bloggers attitude about president changed since election, identifying child suitability of videos based on comments. Although there has been a lot of works published for universal languages like English, works on dialectal languages like Malayalam is comparatively less. But importance of Malayalam is increasing on social medias and shopping sites. This shows the scope of the topic. Another weak point of existing system is that the task done till today is only coarse grained in Malayalam considering only just classification of negative and positive polarity without considering the aspect on which the user is commenting. Such a fine grained task is also considered here most commonly known as Aspect based sentiment analysis. It can contribute to other fields like data mining and web mining.

Keywords: Sentiment Analysis, Aspect-Based Sentiment Analysis, Senti-Wordnet, Polarity, POS tagging.