Artificial Neural Networks (ANN) are mathematical models for the simulation of typical human brain activities such as image perception, pattern recognition, language understanding, sense-motor coordination, and so on. These models are composed of a system of nodes, equivalent to the neurons of a human brain, which are interconnected by weighted links, equivalent to the synapses between neurons. The output of the network is modified iteratively from link weights to convergence. The original data is provided to the input layer and the result of the network is returned from the output level. The input nodes represent the independent or predictor variables that are used to predict the dependent variables, that is, the output neurons.
The MATLAB Neural Network Toolbox provides algorithms, pretrained models, and apps to create, train, visualize, and simulate artificial neural networks. We can perform classification, regression, clustering, dimensionality reduction, time series forecasting, and dynamic system modeling and control. ANNs include data structures and algorithms for learning and classification of data. Through neural network techniques, a program can learn by examples and create an internal structure of rules to classify different inputs.
This chapter shows us how to use neural networks to fit data, to classify patterns, and for clustering. You'll learn preprocessing, postprocessing, and network visualization for improving training efficiency and assessing network performance. We'll also analyze how to properly divide the data before submitting it to the network.
We will cover the following topics:
- Creating, training, and simulating neural networks
- Fitting data
- Using neural network GUI
- Neural network command-line functions
At the end of the chapter, we will understand the basic concepts of artificial neural networks and how to implement them in the MATLAB environment. We will learn how to prepare data for neural network analysis. We will know how to perform data fitting analysis in MATLAB and perform a command-line neural network analysis to understand how specific functions work.