Cicek, GulayBuldag, NazliAydin, Elif2026-01-312026-01-312025979833156656297983315665552165-0608https://doi.org./10.1109/SIU66497.2025.11111864https://hdl.handle.net/20.500.12662/1072233rd Conference on Signal Processing and Communications Applications-SIU-Annual -- JUN 25-28, 2025 -- Istanbul, TURKIYEIn this study, sentiment analysis was performed by creating a dataset of YouTube comments categorized into five emotion classes. Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) were used for feature selection, and classification was conducted using machine learning (Logistic Regression, Random Forest) and deep learning (LSTM, CNN, BiLSTM) algorithms. The highest performance was achieved with the LSTM model on PCA-processed data and with the Logistic Regression model on LDA-processed data, reaching 88% accuracy. This study contributes to the literature with its original dataset and a comparative analysis of the effects of PCA and LDA.trinfo:eu-repo/semantics/closedAccessSentiment AnalysisMachine LearningDeep LearningYouTube CommentsConference Object10.1109/SIU66497.2025.111118642-s2.0-105015521024N/AWOS:001575462500068N/A