Deep Learning Techniques for Social Media Analytics

Abstract

Machine learning has seized both academia and industry's attention as deep learning (DL) is the frontrunner in data science. In order to construct computational models, DL uses multiple layers to epitomize data theories. A few of the key DL techniques, such as model transfer (MT), convolutional neural networks (CNN), and generative adversarial networks (GAN), have completely altered our understanding of information processing. Indeed, DL's processing power while handling images, text, and speech is truly remarkable. Because of the rapid growth and extensive availability of digitized social media (SM), evaluating these data by employing conventional technologies and tools is complex and unmanageable. These challenges are expected to be well managed through solutions offered by DL methods. Hence, we consider the executed DL methods built-in regard to social media analytics (SMA). However, rather than engaging in technical details, we study domains that pose serious challenges to SM where DL is involved and propose solutions to those challenges. We also present a few case studies.