Designing a Neural Network and Its Applications
Common machine learning techniques are used when training and designing a neural network. Neural networks can be classified as:
- Supervised neural networks
- Unsupervised neural networks
Supervised neural networks
These are like the example used in the previous section (predicting the price of the house based on how many rooms it has). Supervised neural networks are trained on datasets consisting of sample inputs with their corresponding outputs. These are suitable for noise classification and making predictions.
There are two types of supervised learning methods:
- Classification
This is for problems that have discrete categories or classes as target outputs, for example the Iris dataset. The neural network learns from sample inputs and outputs how to correctly classify new data.
- Regression
This is for problems that have a range of continuous numerical values as target outputs, like the price of a house example. The neural network...
