Abstract

This research paper aims at explaining the fundamentals of neural network with a very different approach by first taking a real life example to bring simplicity. Also, beginning by explaining a biological neuron which is the real working model of an artificial neuron and diving in to discuss the structural aspects of a neural network along with the activation functions that act as decision makers for an input to a neuron. Further it covers the mathematical aspects of training a neural network by discussing the back propagation algorithm which has been widely accepted as the best known method to learn a neural network. Although a theoretical insight into the concept of neural networks, this paper is a precursor to an implementation of Financial Predictor which is still in progress.