Automatic MOOC Video Classification using Transcript Features and Convolutional Neural Networks

摘要: The amount of MOOC video materials has grown exponentially in recent years. Therefore, their storage and analysis need to be made as fully automated possible order maintain management quality. In this work, we present a method for automatic topic classification videos using speech transcripts convolutional neural networks (CNN). Our works follows: First, recognition is used generate transcripts. Then, the are converted into images statistical co-occurrence transformation that designed. Finally, CNN produce category labels transcript image input. For our data, use Khan Academy on Stick dataset contains 2,545 videos, where each labeled with one or two 13 categories. Experiments show strongly competitive against other methods also based features supervised learning.