ANN may be regarded as an alternative to the standard physical models, particularly in cases where the underlying physics of the system is too complex to analyse. In essence, it is a “black box” model, which mimics the information processing functions of the human neural system. ANN accepts any standard input vector and produces the desired output by processing the input through a series–parallel combination of functional elements, commonly referred as “neurons” or “nodes”. Multilayer perceptron (MLP) neural network is the most widely used ANN architecture. ANN has created newer and massive strides in the field of science.
It has been found to be useful in predicting the survival rate, length of stay in hospitals of patients suffering from trauma or in the intensive care units. ANN being a powerful tool in predicting bivariate models; with recent prediction of the occurrence of heart block and death in patients with myocardial infarctions simultaneously by the use of hybrid models referred to as hybrid ANN-Genetic Algorithm (ANN-GA). ANN has also been successfully used in temperature tracking, constraints and limitations of different products used in summer and winter.
-Deep Neural Network;
-Recurrent Neural Network;
-Dynamic biological networks;
-Adaptive neural processing;