Enhancement of Sentiment Analysis Classification Performance of Spotify Reviews through NLP-based Approaches

Abstract

In this study, sentiment analysis of Spotify reviews was conducted utilizing Natural Language Processing (NLP) techniques. Specific preprocessing steps were implemented to enhance the performance of sentiment classification, aiming to achieve superior outcomes compared to NLP-based solutions. These preprocessing procedures hold significant importance in effectively categorizing emotions. The resultant sentiment analysis categorized sentiments into three classes: negative, neutral, and positive. These derived classes were subjected to a comparative analysis employing machine learning methodologies. The comparative assessment revealed that the Xgboost method exhibited a more successful classification performance compared to other approaches. To address data imbalance within the dataset, various handling imbalance methods were employed and juxtaposed against one another. Through this investigation, the study contributes to a more comprehensive understanding of sentiment analysis of Spotify reviews using NLP, shedding light on optimal preprocessing strategies and effective machine learning techniques for sentiment classification in this context.