ROUGE Score Analysis and Performance Evaluation Between Google T5 and SpaCy for YouTube News Video Summarization

The project aims to compare the performance be-tween Google T5 and SpaCy for YouTube video textual summarization based on the ROUGE score. The project utilizes two distinct techniques: abstractive summarization using Google T5 and extractive summarization using SpaCy. The CNN/DailyMail dataset serves as the primary data source for training the model and the YouTube video description sections are used for evaluating the results. The project evaluates the effectiveness of each technique separately. The strengths and weaknesses of abstractive and extractive summarization approaches are discussed, providing valuable insights into their respective trade-offs and applicability for YouTube news video summarization. From the observations, it is found that Google T5 outperforms SpaCy every single time. The overall ROUGE-1 score for Google T5 is found to be 0.339 while the SpaCy scored 0.232. The findings demonstrate the potential of AI-based techniques for automating the summarization process for YouTube news videos.