Sentiment Analysis Of Spotify App In Playstore Using Classification Method

Abstract

Spotify is a globally renowned music streaming program.  The program receives a multitude of ratings, both favorable and unfavorable, on the Google Play Store from various users.  This study intends to evaluate the sentiment of user evaluations for the Spotify application employing various classification techniques, including Logistic Regression, Random Forest, Support Vector Machine (SVM), C4.5, and Extreme Gradient Boosting (XGBoost).  Review data was acquired via web scraping methodologies using the Google Play Scraper API.  After this, text preparation was conducted to sanitize the text, enabling the execution of the data.  Sentiment analysis was employed to ascertain whether a text expresses favorable or unfavorable opinions.  The Random Forest approach, which has been demonstrated to yield optimal outcomes, was employed in this investigation.  Testing was performed using training and test data ratios of 80:20%, 70:30%, and 60:40% across hundreds of review datasets.  The Random Forest approach, utilizing an 80%:20% data split ratio, produced a precision of 82%, recall of 81%, F1-Score of 81%, and accuracy of 81%, according to the test findings