Performance Evaluation of Machine Learning and Deep Learning Techniques for Sentiment Analysis

摘要

Since the proliferation of opinion-based web content, sentiment analysis as an application of natural language processing has attracted the attention of researchers in the past few years. Lot of development has been brought in this domain that has facilitated in achieving optimal classification of text data. In this paper, we experimented with the widely used traditional classifiers and deep neural networks along with their hybrid combinations to optimize relevant parameters so as to obtain the best possible classification accuracy. We conducted our experiments on labeled movie review corpus and have presented relevant results and comparisons.