VisSim/Neural-Net uses established neural network algorithms to perform learning and recognition of patterns. Neural networks excel 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 neural networks, are used for learning nonlinear dynamic behavior from historic data sets. Once trained, neural networks are used to predict plant behavior based on input values.
Neural networks can be both trained and used for prediction directly from a VisSim diagram.
- Up to 32 layers per network
- Up to 32767 neurons per layer
- up to 128 neural networks per diagram
- Five learning methods
- User-selectable learn rate
- User-selectable neighborhoods
- User-specified smoothing factor
- Table-driven weight initialization