Optimizing YouTube Spam Detection with Ensemble Deep Learning Techniques

A pressing conundrum on social media platforms-Spam comments, have ushered the ecosystem of YouTube suppressing content quality and user experience where recent scams like Ponzi schemes and phishing are hiking. These unnecessary self-promotions and redirection to malicious sites which lead to finance and information loss are aimed to be detected for saving both the content creator's time by making the filtration process easier and also automatic spam removal algorithm more efficient. Although tools like Google Safe Browsing or YouTube team's blocker are in play, they just block basic links and fail to secure users in real-time scenarios. Here we propose a stacking classifier, an ensemble approach towards spam detection that combines various deep learning techniques such as LSTM, CNN, Attention, and a hybrid model of LSTM-CNN to output a classification using a DNN as the meta classifier. Our stacking approach has achieved an accuracy of 93.5% and a comparative study with individual models is also looked into.