Machine Learning and Data Science Blueprints for Finance

Chapter 8. Unsupervised Learning: Clustering

In the previous chapter, we explored dimensionality reduction, which is one type of unsupervised learning. In this chapter, we will explore clustering, a category of unsupervised learning techniques that allows us to discover hidden structures in data.

Both clustering and dimensionality reduction summarize the data. Dimensionality reduction compresses the data by representing it using new, fewer features while still capturing the most relevant information. Similarly, clustering is a way to reduce the volume of data and find patterns. However, it does so by categorizing the original data and not by creating new variables. Clustering algorithms assign observations to subgroups that consist of similar data points. The goal of clustering is to find a natural grouping in data so that items in a given cluster are more similar to each other than to those of different clusters. Clustering serves to better understand the data through the lens of several categories or groups created. It also permits the automatic categorization of new objects according to the learned criteria.

In the field of finance, clustering has been used by traders and investment managers to find homogeneous groups of assets, classes, sectors, and countries based on similar characteristics. Clustering analysis augments trading strategies by providing insights into categories of trading signals. The technique has been used to segment customers or investors into a number of ...