A Classification and Retrieval System for Learning Resources of MOOC

Abstract

In this information age, massive open online courses (MOOCs) have become an integral component of modern education. These courses encompass a wide range of resources, such as videos, audios, texts, and other forms. An accompanying question is how to effectively organize, classify, and retrieve these resources. However, currently available classification and retrieval methods are mostly based on text retrieval technologies. As a result, multi-modal resources such as videos and audios are often ignored or incorrectly classified. Furthermore, more current methods exhibit low efficiency when processing the vast amount of data in MOOCs. To address and solve these issues, this study focuses on the extraction and fusion of multi-modal features of MOOC resources. It proposes an efficient classification and retrieval method based on 3D convolution, aiming to offer a more accurate and efficient approach for classifying and retrieving MOOC resources.