Developing an Innovative NLP based Model to enhance Search Accuracy for Microlearning Videos on YouTube

Abstract

Technology has recently developed to support all fields and increase efficiency, especially in education. Microlearning has emerged as a powerful educational approach, allowing learners to view short, focused content that aligns with their needs and schedules. Micro-learning videos have recently become widespread, but the retrieved videos may not be relevant to the keywords used for search. This research aims to create an educational model specialized in micro-learning using Natural language processing. However, developing an effective microlearning model involves delivering highly relevant content, which requires developing a model that compares two NLP algorithms. Using two advanced NLP algorithms and comparing them after development to choose the best can significantly enhance the framework's accuracy. In this study, we develop and compare the NLTK and Gensim algorithms, measuring their cosine similarity to determine the best. The study's findings confirmed that NLTK outperformed Gensim regarding relevance, clarity, and alignment with the intended learning objectives. To ensure the reliability of the results and achieve high accuracy, we conducted a survey among teachers to select the videos that would be ranked based on their relevance. The survey results are aligned with NLTK's results, underscoring NLTK's potential as a more dependable tool for processing video content in microlearning applications.