|
This paper describes the Neural computing alternative to traditional statistical regression analysis in areas of curve fitting, modeling, and prediction. Several examples are considered describing the applications of neural computing for chemists. Knowledge of mathematics and numerical techniques is not necessary for solving these examples; the network building is not equation-oriented, but is handled by software.
The various steps involved in arriving at the solution to the problem like, network building, learning or training the network, and prediction follow the logical process of "learning" by the human brain.
|
Citation
