Long Short-Term Memory-based Sentiment Classification of Cloud Dataset

Abstract

Text Sentiment Classification is a crucial task for various decision-making processes in many organizations. It identifies the polarity of texts positively and negatively and highlights the opinions and views hidden within the comments or reviews of a product or service. Performing it on big data from social media and related sources is quite tricky and time-consuming. Nowadays, Deep Learning (DL) is widely used for sentiment analysis due to its high performance. In this paper, Recurrent Neural Network (RNN) based Long Short-Term Memory (LSTM) approach is applied to perform sentiment analysis of a cloud review dataset. The cloud review dataset contains cloud consumer reviews regarding the services provided by different cloud service providers and the dataset is achieved as a result of the Harvesting-as-a-Service (HaaS) framework. The study focuses on observing the behaviour of the deep learning RNN-LSTM approach on a cloud dataset. Results of the experiment are evaluated using various evaluation and performance metrics. The approach tends to achieve 95 % accuracy.