Neural Networks

A neural network is basically a mathematical version of what goes on in our brain. The brain consists of neurons, a basic cell, which when interconnected produces what we call a neural network. The sole purpose of a neuron is to receive electrical signals, accumulate them and see if they are strong enough to be passed forward.

A neuron is a cell which accumulates electrical signals. It then compares the accumulated signal with a pre-defined value unique to every neuron, called a bias. Neural networks can be broken down into 3 different layers: input layer, hidden/middle layer, and output layer.

Input Layer: The input layer takes the inputs and passes them to the middle layer. This is its sole task. The input layer never processes data, just passes it along.

Middle/Hidden Layer: This layer is what allows the network to solve complex problems. There can be any number of middle layers but for most tasks, one is sufficient. This layer takes values from the input layer, performs some calculations and passes them to the output layer.

Output Layer: This layer takes input from the middle layer, performs calculations and gives a final result. It is similar to the middle layer but instead of passing values on to the next layer, they are treated as output.

Some practical applications of neural networks include:

• Optical character recognition
• Stock market prediction
• Loan risk analysis
• Image analysis