Semi Supervised Aspect Based Sentiment Analysis Using Artificial Intelligence

Semi Supervised Aspect Based Sentiment Analysis Using Artificial Intelligence

This invention purposes a solution to the above stated problem by providing Contextual Similarity Clusters based on contextual distance of terms in review(s) for a specific category. This is achieved by using Word2Vec Model based on some seed data and, also by mapping explicit & implicit aspects on the cluster based on their contextual distance calculated from Word2Vec. The distance calculated from Word2Vec are added to the most suitable aspect cluster. To find aspects in user’s review(s), extract all possible explicit and implicit terms from review(s). The extracted terms are then mapped on Contextual Similarity Clusters. Finally, create mini-clusters of user’s review(s). After sentence tokenization, map each sentence on mini-clusters. This generates the summary for each mentioned aspect(s) according to user’s review(s).


Leave a Reply

Your email address will not be published. Required fields are marked *