Analyzing Public Perception Toward Artificial Intelligence Educational Content on YouTube Using TF-IDF and Support Vector Machine
Year: 2026
Type: conference-paper
Abstract: YouTube has become a prominent platform for the dissemination of educational content, particularly in the field of Artificial Intelligence (AI). Analysis of user comments on these videos offers valuable insights into public perceptions and attitudes toward AI education. Due to the large volume of comments, manual analysis is impractical. This study examines public perceptions of AI education through sentiment analysis of YouTube comments, using Term Frequency–Inverse Document Frequency (TF-IDF) for feature extraction and Support Vector Machine (SVM) for classification. Comments collected from AI educational channels were preprocessed through data cleaning, case folding, stopword removal, stemming, and class balancing, then explored via Exploratory Data Analysis (EDA) prior to model development. Experimental results show that the TF-IDF–SVM pipeline achieved an accuracy of 79.43%, with precision of 79.76%, recall of 79.43%, and F1-score of 79.37%. Sentiment distribution analysis further revealed that positive sentiment predominated, accounting for 45.2% of audience responses, indicating that AI educational content is predominantly well received. These findings demonstrate the effectiveness of the TF-IDF and SVM combination for sentiment classification on YouTube comments and provide useful insights for content creators and researchers in the AI education domain.
Institution: Binus University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W7204839710