E-Service Quality Analysis for Video Players and Editor App Using Sentiment Analysis and Topic Modeling

Media-sharing networks represent a category of social media platforms that facilitate the dissemination of visual content, particularly videos. The growing importance of video-sharing within the context of user feedback is becoming increasingly evident for organizations, as it offers a highly engaging and immersive experience for viewers. The utilization of videos has become a pervasive medium for social media and communication. This study uses TikTok and YouTube, both prominent video-sharing, to identify the essential future improvements of video-sharing platforms through the analysis of user reviews, employing sentiment analysis and topic modeling techniques. We utilize Naive Bayes classification by initially preprocessing 21,365 reviews and divided into seven categories based on the dimensions of electronic service quality: efficiency, fulfillment, system availability, privacy, responsiveness, compensation, and contact. The accuracy percentages yielded 75.56% and 80.33% for YouTube and TikTok, respectively, showcasing commendable performance in assessing the quality of both platforms. Sentiment analysis identified that both Y ouTube and TikTok have a higher prevalence of negative sentiment. We emphasize three dimensions that garner significant attention when comparing YouTube and TikTok: system availability, fulfillment, and efficiency. The conclusion underscores the satisfactory results obtained from the Naive Bayes classification performance.