Apache Spark 2.x Machine Learning Cookbook

Multilayer perceptron classifier in Apache Spark 2.0

In this recipe, we explore Spark's 2.0 multilayer perceptron classifier (MLPC), which is another name for feed-forward neural networks. We use the iris data set to predict a binary outcome for the feature vectors that describes the input. The key point to remember is that, even though the name sounds a bit complicated, at its core the MLP is just a non-linear classifier for data that cannot be separated via a simple linear line or hyperplane.