Plataforma web de recomendação de vídeos didáticos: desenvolvimento e avaliação de sistema híbrido para estudantes de computação

Abstract

YouTube has established itself as the largest video platform on the Internet, serving as an extensive repository of educational content. This work describes the development and evaluation of a web platform for personalized recommendation of instructional Computer Science videos, motivated by the difficulty students face in finding quality content amidst the large volume available on YouTube. The general objective was to implement a prototype that combines collaborative filtering and content-based techniques through a weighted hybrid approach (0.6 Content-Based Filtering / 0.4 Collaborative Filtering) to improve the accuracy and diversity of suggestions. The methodology included integration with the YouTube Data API, preprocessing of metadata, construction of user profiles, and application of a hybrid recommendation algorithm. The system's effectiveness was evaluated through qualitative user feedback and performance analysis. The results provided support for future enhancements and extensions, showing the feasibility of hybrid solutions in the educational context.