Clustering Web Users with Streaming K-Means

In recent machine learning posts, I've been concentrating on supervised learning algorithms such as Naive Bayes or SVM where you train a model with pre-classified data. This model is then used to predict outcomes on future data. In this post I will be looking at an Unsupervised algorithm called K-Means. K-Means is used to cluster data together useful in a number of situations such as grouping people together. I will be using Spark streaming to show how you can stream events in from a website, run them through a K-Means algorithm and also update the model on the fly.