VisSim/NeuralNet 
VisSim/NeuralNet excels at nonlinear system identification, problem diagnosis,
decision making, prediction, and other problems where pattern recognition
is important and precise computational answers are not readily available.
Within the engineering community, scientists are
using neural networks for learning nonlinear dynamic behaviour from historic
data sets. Once trained, neural networks are used to predict plant behaviour
based on input values.
Neural networks can be both trained and run directly
from a VisSim diagram.
Features
- Back Propagation, Back Propagation with momentum,
LVQ/Kohonen, Probabilistic and General Regression learning methods
- Continuous and discrete outputs
- Supports definition of network topology, learning
coefficients, and methods
- Interactive modification of training characteristics
and learning methods
- Monitor learning error interactively
- Save and restore learned weights
|