Sentiment Analysis from Reviewer's Comments from E-Commerce Sites Using A Natural Language Processing

In the modern world, the developments of textual material are growing exponentially day by day. Instagram and Facebook, in-app messaging services like Telegram and WhatsApp, e-commerce sites like Amazon, Google searches, websites that publish news, and a wide range of resources in generating textual contents. Among these all sources generate enormous amounts of text data every single instant. The analysis of data can aid business owners in assessing the societal impact of their brand, product, or service and subsequently, taking appropriate actions. The core objective of this work is to create consumer review summarization model combining with Natural Language Processing (NLP) and Long Short-Term Memory (LSTM) approaches. For efficiently extracting the features and producing a distinct hybrid feature vector for each review, a hybrid technique combining aspect-related features and review-related features has been given. Sentiment classification is carried out using the LSTM deep learning classifier. Experimental results are useful to find the most wanted products. The performance analysis has been evaluated by different evaluation metrics.