Opinion Mining and Analysis using Deep Learning

Opinion Mining, though seems to be everyday term in the field of NLP, yet has a lot of milestones ahead, waiting to be achieved. In this paper, focus was to find better "Opinion Mining with analyzed sentiments" Model based on accuracy of different deep learning techniques after comparing accuracy levels of different deep learning models on which experiments were done. Recently Deep Learning Techniques have been adopted for this task, as it has been showing very promising results in terms of accurately deducing the emotion behind a text lying in internet database. In this paper, experiments were done with different techniques deep learning has provided us with, in order to create a better Opinion mining model.With the help of the experiments that were conducted, using the well-known IMDb data set, it was found that a combination of carefully curated model of selective layers can produce better results than just adding on CNN layers in hopes of getting results. The model is then tested on other datasets and is found to be better than other similar state-of-the-art models.