Volume 18, No. 6, 2021
Big Data Analytics For Sentiment Analysis From The Reviews And Feedbacks Of E-Commerce Portal
Chetan Kumar Soni and Atul D. Newase
Abstract
Recently, big data analytics has emerged as boon in the field of computing technology which is capable of dealing with the very huge amount of data of varying velocity and variety in a very efficient way. The parallel distributed computing architecture with high computing capabilities has made it possible to access, process and analyze this big data. This paperpresents an implementation of this big data analytics technology to analyze the text data received from reviews and the feedbacks of the users of various e commerce companies and identify the sentiments of the users about the specific product or service. The concept of natural language processing is used in this work to identify the orientation of the users. The sentiments are classified into four categories in this work as positive, negative, neutral and unsure. A dictionary is prepared on the basis of keywords projecting the sentiments and the emotions of the users which is preprocessed as per the requirement. The novelty of this work lies into the deep learning algorithm presented in this work which classifies the emotions of the users into the above mentioned four categories and presents a complete framework for the stakeholders to take the necessary action. A neural network based deep learning algorithm is presented in this work and the implementation is done through Hadoop. As compared to the previously proposed algorithms, this model presents a larger range of emotions over a huge database. It also provides a better accuracy and efficiency.
Pages: 6884-6895
Keywords: Sentiment analysis, Emotion mining, Deep learning, Neural networks, Big data analytics, Hadoop.