A Multi-modal Approach of Sentiment Analysis on Samsung Cell Phone Review

Sentiment analysis, the scientific art of understanding human emotions. The motive of sentiment analysis is to extract and predict human emotions and sentiment through text, speech or facial expressions. It is proposed an integrated study that uses text review given by buyers of Samsung phones at Amazon to analyze sentiments and satisfaction towards a Samsung phone. In research, it has been seen that potential customers look for text reviews besides numeric rating to get an overview of positive and negative impressions of the product. This research helps to get critical insights for analyzing customers’ text review and giving a sentiment score for a specific text review by sentiment analysis algorithmic model. This research compares numeric and text reviews based on sentiment scores and predicts overall customer sentiment based on a text review’s sentiment score. The objective of this research is to analyze customer’s sentiment on Samsung cell phone using a multimodal approach (TextBlob, VADER, and LSTM). It extracted positive and negative features of Samsung brand or product as per buyers’ opinion, which is called opinion mining using Word Cloud dictionary library of NLP. Applied LDA model for topic modeling, used XGBoost to extract feature importance for sentiment classification, implemented VADER and TextBlob algorithm for sentiment analysis, and tested accuracy to VADER and TextBlob, applying LSTM for sentiment prediction.The research’s implication is that both customers and brand understand the recent market trends directly through the customer’s sentiment. Customers can decide to buy or not a product; the brand can find out where to improve a product.