Sentiment Analysis of YouTube Comments on AI-Mediated Digital Resurrection

This dataset contains YouTube comments collected to analyze public sentiment towards AI recreation of deceased individuals. The data were obtained using the YouTube API and subsequently processed through several stages, including data cleaning, language filtering to retain English comments, sentiment labeling using the VADER lexicon-based method, and text preprocessing for machine learning purposes. A hybrid sentiment analysis approach was implemented, combining lexicon-based sentiment scoring (VADER) with a machine learning classifier, specifically Support Vector Machine (SVM) with TF-IDF feature extraction. The dataset includes multiple versions of the data, namely the raw comments, cleaned dataset, English-filtered dataset, VADER-labeled dataset, preprocessed dataset, and the final dataset prepared for SVM training. This dataset is provided to support transparency and reproducibility of the research.