Artificial Neural Networks
Neural networks — the “artificial” is usually dropped — are a class of powerful, flexible, general-purpose techniques readily applied to prediction, estimation, and classification problems. Applications include detecting fraudulent credit card transactions, modeling financial time series, recognizing handwritten numbers and letters, and estimating real estate values.
The first artificial neural networks were conscious attempts to simulate the workings of biological neural networks using digital computers. In addition to biologists interested in the workings of the nervous system, early artificial intelligence researchers saw neural networks as a way to endow computers with the ability to learn. The human brain makes generalizing from experience possible for people; computers, on the other hand, usually excel at following explicit instructions over and over. The appeal of neural networks is that they bridge this gap by modeling, on a digital computer, the neural connections in human brains. When used in well-defined domains, their ability to generalize and learn from data mimics, in some sense, the human ability to learn from experience.
Because of this history, neural networks researchers originally used terminology from biology and machine learning that was quite dissimilar to the terminology used in statistical modeling; it took some time before neural networks were recognized as a useful modeling method by statisticians. In the past twenty ...