Sentiment Analysis for Product Reviews Based on Weakly-Supervised Deep Embedding

Abstract

Buyers to whom a product would be introduced should check it to make better choices about the item. To arrive at a specific finding, various viewpoint mining methods have been suggested. Several recent developments in machine learning, especially deep learning, have led to considerable progress in solving sentiment classification problems. To achieve valuable scores as poor supervision indicators, this research work suggests an innovative deep learning system for performing product review based emotion classification. To achieve a high-level representation, one needs to learn the embedding before applying a classification layer on top of the embedding.

Topics & Concepts