A review of online supervised learning

Online learning emerged as a promising solution to handle large data problems. The high-level performance witnessed in real-life applications of online learning established dramatic advances in this field. The varying nature of the data needs special attention from a research point of view, as it has emerged as a common challenge in many domains. Interestingly, online learning response to this varying nature of the data is one of the promising solutions. We continue in this direction by covering successful algorithms in literature and their complexities to meet new challenges in this field. In particular, we have covered the working of online supervised learning algorithms and their bounds on mistake rate. A suitable and systematic review of online supervised learning algorithms is crucial for domain understanding and a step toward a solution to meet future challenges in this field. We have covered online supervised learning review with its common framework, algorithms description in ascending order of their development of applications in real-life use, and discussion on their theoretical analysis of algorithms. The present paper also includes an experimental comparison to understand advances in online learning algorithms responses to benchmarked datasets as well as future challenges in this field.