|
 |
|
|
|
|
|
|
|
NeuroSolutions Products Level Compare
|
NeuroSolutions is broken down into three different levels which cater for different capabilities. You only have to pay for the capabilities that are useful to you. A - STANDARD, B - PROFESSIONAL, C - DEVELOPERS |
|---|
|
NeuroSolutions Level Summary |
A | B | C | |||
|---|---|---|---|---|---|---|
| Topologies | ||||||
| Multilayer Perceptron (MLP) | Y |
Y |
Y |
|||
| Generalized Feedforward Network | Y |
Y |
Y |
|||
| Modular Network | Y |
Y |
Y |
|||
| Jordan / Elman Networks | Y |
Y |
Y |
|||
| Self-Organizing Map (SOM) | Y |
Y |
Y |
|||
| Principal Component Analysis (PCA) | Y |
Y |
Y |
|||
| Radial Basis Function (RBF) | Y |
Y |
Y |
|||
| Probabilistic Neural Network (PNN) | Y |
Y |
Y |
|||
| General Regression Neural Network (GRNN) | Y |
Y |
Y |
|||
| Neuro-Fuzzy Network (CANFIS) | Y |
Y |
Y |
|||
| Support Vector Machine Network | Y |
Y |
Y |
|||
| Hopfield Network | Y |
Y |
||||
| Time Delay Neural Network (TDNN) | Y |
Y |
||||
| Time-Lag Recurrent Network (TLRN) | Y |
Y |
||||
| General Recurrent Network | Y |
Y |
||||
| Maximum Number of Inputs / Outputs / Neurons Per Layer | 500 |
Unlimited |
||||
| Maximum Number of Hidden Layers | 6 |
Unlimited |
||||
| Learning Paradigms | ||||||
| Backpropagation | Y |
Y |
Y |
|||
| Unsupervised Learning (Hebbian,Ojas,Sangers,Competitive,Kohonen) | Y |
Y |
Y |
|||
| Recurrent Backpropagation | Y |
Y |
||||
| Backpropagation through time | Y |
Y |
||||
| Gradient Descent Methods | ||||||
| Step / Momentum | Y |
Y |
Y |
|||
| Delta Bar Delta | Y |
Y |
Y |
|||
| Quickprop | Y |
Y |
Y |
|||
| Conjugate Gradient | Y |
Y |
Y |
|||
| Levenberg-Marquardt | Y |
Y |
Y |
|||
| Advanced Features | ||||||
| Exemplar Weighting - Improved training for data with unequal class distribution | Y |
Y |
Y |
|||
| Macros / OLE Automation - API to Automate and control NeuroSolutions | Y |
Y |
Y |
|||
| Sensitivity Analysis - to determine the most influential inputs | Y |
Y |
Y |
|||
| Genetic Optimization - Intelligent searching for the best parameters and inputs | Y |
Y |
Y |
|||
| Iterative Prediction - Advanced method for time series prediction | Y |
Y |
||||
| ANSI C++ Source Code Generation - Embedded neural networks into your own applications, train neural networks on faster computers | Y |
Y |
||||
| User-defined Neural Components (using DLLs) (Nonlinearities,Interconnections, Learning Rules, Error Criteria, input/Output, Memory Structures | Y |
|||||
 |
 |