Neural Network

Artificial_neural_network

artificial neural network (ANN) is the network of a group of small processing units that are modeled on the human nervous tissue. ANN is an adaptive system which can change its structure to solve problems based on external or internal information that flows through the network.
Simply put, ANN is a statistical data modeling tools of non-linear. ANN can be used to model the complex relationships between inputs and outputs for finding patterns in data.

History

Currently the field of artificial intelligence in his attempt to imitate human intelligence, has not been entered in the form of his physical approach but from the other side. First conducted a study on the basic theory of the mechanism of the intelligence process. This field is called ‘Cognitive Science’. From this basic theory is made of a model to be simulated on a computer, and in the further development of artificial intelligence known to the various systems one of which is an artificial neural network. Compared with other fields of science, artificial neural networks is still relatively new. A number of literature considers that the concept of neural networks began in the Waffen paper McCulloch and Walter Pitts in 1943. In this paper they are trying to formulate a mathematical model of brain cells. The method was developed based on this biological nervous system, represents a step forward in the computer industry.

Basic Definition

No two human brains, each brain is always different. Differences in sharpness, size and organized. One way to understand how the brain works is to gather information from as many as possible of human brain scanning and mapping. It was an attempt to discover how the average human brain. Map of the human brain are expected to explain the mystery of how the brain controls every human action, ranging from the use of language to the movement.
Nevertheless certainty how the human brain is still a mystery. Although several aspects of this amazing processor are known but it is not a lot. Some of these aspects, namely:
a. Each part of the human brain has an address, in the form of chemical formula, and the human nervous system trying to obtain a suitable address for each axon (nerve liaison) established.
b. Through learning, experience and interaction between the systems of the brain structure itself will regulate the functions of each part.
c. Axon-Axon in the adjacent areas will evolve and have a similar physical form, so clustered with a specific architecture in the brain.
d. Axon growth in the order based on the architecture of time, and connects to the brain structure that develops with the same timeline.
Based on the four aspects mentioned above can be drawn a conclusion that the brain is not entirely formed by the genetic process. There are other processes that help shape the function of parts of the brain, which in turn determines how the information is processed by the brain.
The most basic element of neural networks are the nerve cells. These nerve cells form part of human consciousness that includes some general ability. Basically the biology of nerve cells receive input from other sources, and combine them with some way, performs a non-linear operations to get results and then issued the final results.
In the human body there are many variations of the basic types of nerve cells, so that the human thinking process becomes difficult to be replicated electrically. Notwithstanding the foregoing, all natural neurons have the same four basic components. The four basic components are known by the name of biology is, the dendrites, soma, axons, and synapses. Dendrite is an extension of the soma-like hair and act as a channel input. These input channels receive input from other neurons via synapses. Soma in this matter and then process the input into an output value is then sent to other neurons through axons and synapses.
Recent research provides further evidence that biological neurons have more complex structures and more sophisticated than the artificial nerve cells which then formed into an artificial neural network, there is now. Biological sciences to provide a better understanding of nerve cells that give the advantage to the network designer to be able to continue to improve artificial neural network system that is based on the understanding of brain biology.
Nerve cells, nerve cells are connected to each other through synapses. Nerve cell can receive a stimulus in the form of electrochemical signals from nerve cells to other nerve cells. Based on these stimuli, nerve cell will send a signal or not based on certain conditions. The basic concept of this kind who want to try the experts in creating artificial cells.

Definition

An artificial neural network to process vast amounts of information in parallel and distributed, it is inspired by models of biological brains work. Some definitions of neural networks is as follows below.
Hecht-Nielsend (1988) defines artificial neural systems following:
“A neural network (NN), is a distributed information processing structure and work in parallel, which consists of processing elements (which has a local memory and operating with local information) that diinterkoneksi together with unidirectional signal flow, called connections. Each processing element has a single output connection that branches (fan out) to the desired number of collateral connections (each connection carrying the same signal from the output of these processing elements). The output of these processing elements can be any desired type of mathematical equation. The whole process that goes on every processing elements should really be done locally, ie the output depends only on the input value at that time obtained through the connection and the value stored in local memory. ”
According to Haykin, S. (1994), Neural Networks A Comprehensive the Foundation, NY Macmillan, neural network defines as follows:
“A neural network is a parallel and distributed processors mempuyai tendency to store the knowledge acquired from experience and make it still be available for use. This resembles the workings of the brain in two respects, namely: 1. Knowledge obtained by the network through a learning process. 2. Strength of the relationship between nerve cells, known as synaptic weights are used to store knowledge.
And according Zurada, J.M. (1992), Introduction To Artificial Neural Systems, Boston: PWS Publishing Company, defines as follows:
“Neural system or neural network is a physical cellular systems which can acquire, store and use the knowledge gained from experience.”
DARPA Neural Network Study (1988, AFCEA International Press, p. 60) defining the artificial neural network as follows:
A neural network was a system that is formed from a number of simple processing elements working in parallel in which the function is determined by network structure, power relationships, and pegolahan performed at computing elements or nodes

Keyword :