Finding Groups of Data – Clustering with k-means
Have you ever spent time watching a crowd? If so, you have likely seen some recurring personalities. Perhaps a certain type of person, identified by a freshly pressed suit and a briefcase, comes to typify the “fat cat” business executive. A 20-something wearing skinny jeans, a flannel shirt, and sunglasses might be dubbed a “hipster,” while a woman unloading children from a minivan may be labeled a “soccer mom.”
Of course, these types of stereotypes are dangerous to apply to individuals, as no two people are exactly alike. Yet, understood as a way to describe a collective, the labels capture some underlying aspect of similarity shared among the individuals within the group.
As you will soon learn, the act of clustering, or spotting patterns in data, is not much different from spotting patterns in groups of people. This chapter describes:
- The ways clustering tasks differ...
