Chapter 2. Unsupervised Machine Learning

As we have seen in the Chapter 1, Introduction to Practical Machine Learning Using Python, unsupervised learning is designed to provide insightful information on data unlabeled date. In many cases, a large dataset (both in terms of number of points and number of features) is unstructured and does not present any information at first sight, so these techniques are used to highlight hidden structures on data (clustering) or to reduce its complexity without losing relevant information (dimensionality reduction). This chapter will focus on the main clustering algorithms (the first part of the chapter) and dimensionality reduction methods (the second part of the chapter). The differences and advantages of the methods will be highlighted by providing a practical example using Python libraries. All of the code will be available on the author's GitHub profile, in the https://github.com/ai2010/machine_learning_for_the_web/tree/master/chapter_2/ folder. We will now start describing clustering algorithms.