August 23, 2017
K-means is a clustering algorithm used to categorize data that is unlabeled. It is a type of unsupervised learning (data without categories or groups). The goal is simply to find the groups in the data defined by K. The algorithm works by iteratively assigning each data point to one of the K groups based on the features provided.
Some of the use cases for K-Means are:
In this step we need to find where the new centroid for each cluster is located. We can do so by imagining the current centroids are weightless but no the other points. We can then find the center of gravity (where most of the points are located) to pin point the new centroid.
Sr. Machine Learning Engineer @Pluralsight. Interested in distributed systems, machine learning, and the web. Follow me on Twitter.