This concludes not only the journey inside the multilayer perceptron, but also the introduction of the supervised learning algorithms. In this chapter, you learned:
The components and architecture of a neural networks
The stages of the training cycle of a backpropagation multilayer perceptron
How to implement an MLP from the ground up in Scala
The numerous configuration parameters and options to use MLP as a classifier and regression
To evaluate the impact of the learning rate and the gradient descent momentum factor on the convergence of the sum of squared errors during training
How to apply a multilayer perceptron to the financial analysis of the fluctuation of currencies
The next chapter will introduce the concept of genetic algorithms with a full implementation in Scala. Although, strictly speaking, genetic algorithms do not belong to the family of machine learning algorithms, they play a crucial role in the optimization of nonlinear, nondifferentiable problems and the selection of strong classifiers within ensembles.