Sentiment Analysis on YouTube Smart Phone Unboxing Video Reviews in Sri Lanka

TL;DR

This study analyzes YouTube unboxing videos of three 2021 smartphones in Sri Lanka, using sentiment analysis to gauge user feedback and compare classifier performances, highlighting positive sentiments especially for iPhone 13.

Contribution

It applies lexicon-based sentiment analysis to YouTube reviews of smartphones in Sri Lanka and evaluates machine learning classifiers for sentiment classification accuracy.

Findings

All three smartphones received predominantly positive reviews.

Support Vector Machine outperformed other classifiers in accuracy.

iPhone 13 had the highest number of positive reviews.

Abstract

Product-related reviews are based on users' experiences that are mostly shared on videos in YouTube. It is the second most popular website globally in 2021. People prefer to watch videos on recently released products prior to purchasing, in order to gather overall feedback and make worthy decisions. These videos are created by vloggers who are enthusiastic about technical materials and feedback is usually placed by experienced users of the product or its brand. Analyzing the sentiment of the user reviews gives useful insights into the product in general. This study is focused on three smartphone reviews, namely, Apple iPhone 13, Google Pixel 6, and Samsung Galaxy S21 which were released in 2021. VADER, which is a lexicon and rule-based sentiment analysis tool was used to classify each comment to its appropriate positive or negative orientation. All three smartphones show a positive sentiment from the users' perspective and iPhone 13 has the highest number of positive reviews. The resulting models have been tested using N\"aive Bayes, Decision Tree, and Support Vector Machine. Among these three classifiers, Support Vector Machine shows higher accuracies and F1-scores.