An Approach Towards Enhancement of Review Summarization Using Sequence Model And Word Embedding

Traditionally, extraction-based text summarization techniques have simply relied upon statistical measures for extraction of major points in a document. With popularization of deep learning in solving problems in the field of Natural Language Processing, there is a need to incorporate deep learning to improve the effectiveness of summarization methods. In this paper, we explore the results and findings of our approach to improve review summarization as well as to delineate the capabilities of these language models in improving Natural Language Understanding. The aim of our summarization approach is to generate a gist of customers' feedback of a product which would provide vendors and other customers with the necessary understanding in order to make informed purchases. Our proposed summarization approach is split into four major steps: Keyword Extraction, Sentiment Analysis, Adjective Mining and Sentence Extraction. Afterwards, all the outputs of these steps are combined to provide the user with summary of all reviews about a product.