<![CDATA[Sentiment analysis of comments on YouTube videos related to MPL Season 13 was conducted using an Ensemble Learning-based classification method. This study focuses on identifying sentiment patterns in comments and determining team popularity based on positive fan support. The methods employed include initial planning, data collection through scraping techniques using the YouTube Data API v3, and preprocessing steps. From a total of 6,424 comments collected, the number of relevant comments was reduced to 5,185 after the cleaning, case folding, stopword removal, slang conversion, stemming, and tokenization stages, resulting in 3,131 positive comments and 2,064 negative comments. Various classification methods were applied simultaneously and combined using ensemble machine learning techniques with a majority voting approach. Before classification, the data was processed using SMOTE (Synthetic Minority Over-sampling Technique) to address class imbalance. The testing results showed that the hard voting method achieved an accuracy of 86,70% (with 90% training data and 10% testing data), while the soft voting method reached an accuracy of 86,17%. The labeling process was carried out using the Flair library, validated by a confusion matrix. The application of a labeling method that combines both automatic and manual approaches successfully improved classification accuracy and minimized potential errors. Additionally, this analysis identified the highest supporter count, with 877 supporters for EVOS, followed by RRQ and ONIC with 743 and 556 supporters, respectively. This research is expected to make a significant contribution to the development of sentiment analysis in the context of e-sports and open up opportunities for further analysis in future research.]]>
