Introduction
As we are moving towards a revolutionary era of technology so there is a requirement of precise, predictive and accurate analysis which may be fulfill by neural networks.
Neural networks, as the name suggests, is analogous to human brain which have highly interconnected neurons. Different Combinations of transistors formed neurons which are interconnected to design neural networks. These neurons are also termed as processing elements.
Neural networks can efficiently solve a complex problem. By understanding the relationship between the different input data theoretically unknown solution can be found in problem space. Generally the output of neural networks is not predictive.
Now let us take an example of neural networks i.e. “winner takes all circuit”. In WTA circuits many inputs are given to the different neurons and these neurons compete with each other and give output according to neuron having highest input current/voltage.
• Pattern recognition : neural networks play a vital role in pattern recognition. Networks is design according to the input pattern and it gives the output corresponding to associated pattern
• Data validation : several data is given as input at different layers of neuron but result come out according to the input data validate by neural circuitry.
• Forecasting : By analyzing the output at different inputs forecasting can be done i.e. it can be determine that which input gives desired output.
• Target marketing : if forecasting can be done then way to achieve target can be analyzed easily. It becomes possible to know that which input leads target output.
• Risk management : with the help of neural networks chances to choose wrong input becomes quite less.
As we are moving towards a revolutionary era of technology so there is a requirement of precise, predictive and accurate analysis which may be fulfill by neural networks.
Neural networks, as the name suggests, is analogous to human brain which have highly interconnected neurons. Different Combinations of transistors formed neurons which are interconnected to design neural networks. These neurons are also termed as processing elements.
Neural networks can efficiently solve a complex problem. By understanding the relationship between the different input data theoretically unknown solution can be found in problem space. Generally the output of neural networks is not predictive.
Now let us take an example of neural networks i.e. “winner takes all circuit”. In WTA circuits many inputs are given to the different neurons and these neurons compete with each other and give output according to neuron having highest input current/voltage.
Application of Neural Networks
There are various applications of neural networks in today’s industries, some of them are as follows:
• Pattern recognition : neural networks play a vital role in pattern recognition. Networks is design according to the input pattern and it gives the output corresponding to associated pattern
• Data validation : several data is given as input at different layers of neuron but result come out according to the input data validate by neural circuitry.
• Forecasting : By analyzing the output at different inputs forecasting can be done i.e. it can be determine that which input gives desired output.
• Target marketing : if forecasting can be done then way to achieve target can be analyzed easily. It becomes possible to know that which input leads target output.
• Risk management : with the help of neural networks chances to choose wrong input becomes quite less.
 |
| Neural Networks |
Author - Jyotishna Bansal
(Intern at Silicon Mentor)
No comments:
Post a Comment