Machine Learning vis-à-vis Mathematical Modeling

Having learned about the main components of mathematical optimization, which are decision variables, objective functions, and constraints, in the previous chapter, it is time to throw light on machine learning (ML) models, most of which can be cast as mathematical models. Humans make machines learn from huge amounts of historical data. ML models enhance the decision-making abilities of man and machine, exploiting the power of data and algorithms. There is almost always some optimization algorithm working in the background of most of these models.

The term ML was first popularized by Arthur L. Samuel in the 1950s, who was a pioneer in computer science and gaming. Data volume has increased by leaps and bounds since then, particularly in the last couple of decades, and making sense of huge amounts of data is beyond the scope of the human mind. Hence, ML stepped in and found its application in almost all domains to assist humans...