The Role of Machine Learning in Data Analytics: A Review of Unsupervised Learning Algorithms

Abstract

In this chapter, we will discuss unsupervised learning in general terms. An individual engages in unsupervised learning when they have only access to the input data and do not have access to the corresponding output variables. Unsupervised learning has as its goal the manipulation of the underlying structure and distribution of data to better understand it. Because, unlike supervised learning, which was previously discussed on this thread, there are no correct answers and no instructor present, these are referred to as unsupervised learning. To discover and present the interesting structure within the data, algorithms are left to their own devices. Unsupervised learning problems are further divided into two types: clustering problems and association problems. Clustering problems are the most common type of unsupervised learning challenge.