Transformer-based modelling for sentiment analysis of coffee review

Abstract

The development of social networks contributes to the growth of online content that becomes more relevant to the business. Customer reviews, feedback, and comments have been found mentioned online. Thus, sentiment analysis becomes a tool for businesses to extract insightful information from big data regarding customers’ needs and perceptions towards brands. In this study, the researcher aimed to set the experiments using an open dataset (Yelp Coffee Review) from kaggle.com as a case study. The natural language processing model was implemented by applying the Transformer-based models: BERT, ALBERT and RoBERTa, with six different classifications, including Logistic Regression, Naive Bayes, Decision Tree, Random Forest, Support Vector Machine, and K-Nearest Neighbors, to compare the performance, where imbalance data and hyperparameters have been considered in this study. In the experimental result, the research found that the RoBERTa model with Random Forest classifier performs best at 98.84% accuracy, while all the Transformer-based models outperform the conventional models as expected. Therefore, the models have proved their efficiency and advantage in shedding light on the customer sentiment analysis tool for business. Implication of the thematic paper: The findings of the study indicate that Transformer-based models offer businesses significant advantages and efficiencies for sentiment analysis by accurately analyzing customer sentiment data. The insightful data could support businesses in terms of decision-making, improving customer experiences, and optimizing marketing strategies.