An autonomous system designed for automatic detection and rating of film reviews. Extraction and linguistic analysis of sentiments.

Abstract : This paper describes the functions of a system designed for the assessment of movie reviews. Such a system enables the automatic collection, evaluation and rating of film critics' opinions of movies. First the system searches and retrieves probable movie reviews from the Internet, especially those expressed by prolific reviewers. Subsequently the system carries out an evaluation and rating of those movie reviews. Finally the system automatically associates a numerical mark to each review, this is the objective of the system. This data constitutes the input to the cognitive engine. Our system uses three different methods for classifying opinions in critics' reviews. We introduce two new methods based on linguistic knowledge. Results are then compared with the overall statistical method using Bays classifier. The last step is to combine the results obtained in order to make the final assessment as accurately as possible.