Sentimental Analysis of Amazon Reviews using Map Reducer Technique in Hadoop

Sentiment analysis (SA) is an active and dynamic aspect of text mining that focuses on the automated analysis of subjectivity, opinions, and sentiments in textual material. The study explores new SA domains such as resource development, emotion recognition, and transfer learning. The study combines rule-based classification, machine learning, and supervised learning in a novel hybrid technique. The usefulness of this hybrid method is demonstrated through experiments on multiple datasets, including MySpace comments, product reviews, and movie reviews, resulting in improved classification efficacy as judged by both macro and micro-averaged criteria. The study also supports an additional semi-automatic approach in which classifiers cooperate to enhance overall performance. In this article, we have applied SVM, Decision tree, and Linear SVM to the Amazon review dataset and got the output in the form of performance metrics such as CPU-time and maximum memory usage. and we found that Depending on the linearity of the data, a linear SVM or a Decision Tree is more appropriate; select a linear SVM for clearly defined separations and a Decision Tree for intricate, non-linear interactions, customizing the model to the features of the dataset and the goals of the research.