Building a Clustering Model with Spark - Machine Learning with Spark - page 272

Training a clustering model
Training for K-means in MLlib takes an approach similar to the other models—we pass an
RDD that contains our training data to the train method of the KMeans object. Note that
here we do not use LabeledPoint instances, as the labels are not used in clustering;
they are used only in the feature vectors. Thus, we use a RDD [Vector] as input to the
train method.