PENDEKATAN HIBRIDA PERINGKASAN TRANSKRIPSI VIDEO DENGAN KLASTERISASI K-MEANS BERBASIS BERT DAN BART

The use of video as a medium for delivering information and education continues to grow across online platforms. However, long durations and unstructured delivery often make it difficult for viewers to grasp the main points, creating challenges for developing automatic summarization methods for monologues, interviews, and podcasts. Extractive methods tend to produce less cohesive summaries, while abstractive methods risk omitting important details. To address these issues, this study proposes a hybrid approach that integrates extractive and abstractive techniques. In the extractive stage, sentences are represented using BERT embeddings and grouped using the K-Means Clustering method. The abstractive stage then employs the BART model to generate summaries that are more coherent and informative. Experimental evaluations on 20 Human Metapneumovirus (HMPV) videos show the strongest performance for monologue-type content, achieving ROUGE-1 of 57%, ROUGE-2 of 30%, and ROUGE-L of 32%. Although lower performance was observed for interviews and podcasts due to dynamic interactions and frequent speaker shifts, the hybrid approach consistently outperformed both extractive-only and abstractive-only baselines. These findings highlight the potential of the hybrid approach in producing more adaptive video summarization in the future.