The base of the ANN is the individual nodes which are composed artificially to represent the behavior of a neurons inside human body. The group or a collection of such a artificial nodes is called as a ANN.
Human biological neural network is a complex element which is used to carry out a specific physiological function. So do the Artificial Neural Network. ANN is also designed as the need of it, to perform the function which is intended to.
The flow of a ANN is it captures an input and then using the set of neural nodes, which is called as the ANN, it classifies the input into different classes and give it away as the output. There are main two types of ANNs based on how the training is carried out which is the most important part in any neural network. Training is of two types, supervised training and unsupervised training. In supervised training weights of each node is adjusted in such a way that the output required is obtained. But in unsupervised learning ANN is allowed to train on it's own. Desired output is not provided. It adjusts according to the changes which is more flexible.
When ANNs are used in applications there may be compromises. Since complicated ANNs use huge amount of memory, there could be a clash between quality and feasibility. It should be properly handled in order to produce a good output.
In many of the applications today, they uses neural networks. Some of they would be Function approximation, regression analysis, time series predictions, classification, pattern recognition, time series analysis, data processing.
No comments:
Post a Comment