Abstract:
Objective-Our objective is an attempt to do aspect based sentiment analysis in Malayalam. The main problems existing in this domain are there only coarse grained tasks considering only polarity, but not fine grained tasks considering the aspect on which the user is commenting. Although so many works were proposed for universal languages like English, it is less for dialectal Malayalam.
Design / Methodology/ Approach- To develop a movie site which consist of a database contain the whole details of movie. The sentiment words are inputted by the user, the system will generate appropriate response according to the input. Result can be displayed in the form of emotions with emoji’s.
Findings- It consist of three sections. Admin consist of whole details of registered users and movie reviews. Admin can login to the movie site and add/remove specified informations. Next, the User login to the movie site, input the review and submit it. The system will generate corresponding output and display the result using emotions and emoji’s. Guest, the non registered user can visit the movie site and understand the overall review of the movies.
Limitations- It does not consider a large passage as the input.
Practical implications- This paper is very helpful for those who find sentiment analysis in Malayalam in a movie site.
Originality/Value- This will help to express the emotions in Malayalam terms with emoji’s.
Keywords- Natural Language Processing, Sentiment Analysis, Aspect-Based Sentiment Analysis, Polarity, Senti-Wordnet.