RecoMob : Opinion Mining for Product Enhancement

Thursday. March 31, 2016 - 1 min

You can get code Here Paper can be found Here —

The project was aimed at helping the manufactures to improve their products by getting aspect based sentiment from reviews.

• Used Recursive Neural Tensor Network with Word Vectors.
• Used StanfordCoreNLP toolkit and its dependency parsing framework to extract the aspects and sentiment associated with them by formulating rules.
• Also analyzed how reviews of the product fare over the time.
• Used a k-NN for recommending products to the user based on specific aspects.