Abstractive Summarization Through Sentiment Analysis Of User Product Reviews - An RNN Approach

摘要

This research study focuses on performing granular sentiment detection, using a recurrent neural network approach, over a dataset of online reviews for Slack, an American cloud-based set of proprietary team collaboration tools and services. A client interface was built, attaching sentiment scores to opinion units, keywords and performance indicators, and generating interactive visualizations of the same. An encoder-decoder model with Attention was also implemented, to generate an abstractive customer feedback summary for the client, as a premise for future product development.