Enhancing YouTube Spam Filtration Efficiency Through Deep Learning Based Techniques

In this paper Deep Learning (DL) based techniques for filtering YouTube comments are explored. This work primarily focuses on the applications of Feedforward Neural Networks (FNNs) and Recurrent Neural Networks (RNNs) to increase the accuracy of spam detection. The conventional ML based techniques are comparatively less efficient than DL based approach while handling the complexity of natural language in order to learn the hierarchical representations. This study covers the model architecture and training of FNNs and RNNs. The experimental results represent the ability in accurate classification of YouTube comments. The results achieved in this work shows that Deep Learning models are more efficient with increased accuracy in spam detection over huge datasets.