Exploring artificial neural network hyperparameters
An artificial neural network, also known as deep learning, is a kind of ML algorithm that mimics how human brains work. Deep learning can be utilized for both regression and classification tasks. One of the main selling points of this model is its ability to perform feature engineering and selection automatically from the raw data. In general, to ensure this algorithm works decently, we need a large amount of training data to be fed to the model. The simplest form of a neural network is called a perceptron (see Figure 11.4). A perceptron is just a linear combination that is applied on top of all of the features, with bias added at the end of the calculation:
Figure 11.4 – Perceptron
If the output from the perceptron is passed to a non-linear function, which is usually called an activation function, and then passed to another perceptron, then we can call this a multi-layer perceptron (MLP) with one layer...