NeuroSolutions is broken down into three different levels. This way you only have to pay for the capabilities that are useful to you. Below is a summary of these levels. If any of this information is unclear, please do not hesitate to contact our sales staff -- they would be glad to recommend a level that best meets your needs.
The Standard is the entry level of NeuroSolutions and it is intended for those who want to learn about neural networks. However, this level can be used to solve real-world problems with the most common neural network architecture, the multilayer perceptron (MLP).
Topologies
Multilayer Perceptron (MLP)
Generalized Feedforward
User-defined network topologies
Modular
Jordan/Elman
Self-Organizing Map (SOM)
Principal Component Analysis (PCA)
Radial Basis Function (RBF)
Probabilistic Neural Network (PNN)
General Regression Neural Network (GRNN)
Fuzzy Logic networks
Support Vector Machine networks
Up to 500 inputs/outputs/neurons per layer
Up to 6 hidden layers
Learning Paradigms
Backpropagation
Step
Momentum
Delta-Bar-Delta
Quickprop
Conjugate Gradient
Levenberg-Marquardt
Unsupervised Learning
Hebbian
Oja's
Sanger's
Competitive
Kohonen
Additional Features
Genetic optimization of neural network parameters and inputs
Competitive Advantage
Double-precision calculations
Faster simulations
Icon-based graphical user interface
Extensive probing capabilities
Hybrid supervised/unsupervised topologies
Level B - Professional
The Professional level offers enhanced models which support dynamic pattern recognition, time-series prediction and process control problems. It gives you access to all available neural components and does not restrict the number of inputs, outputs or hidden neurons.
ANSI-compatible C++ code generation, allowing you to embed NeuroSolutions' algorithms into your own applications (including learning). Furthermore, this level allows any simulation prototyped within NeuroSolutions to be run on other platforms (requires the Source Code License). Another option is to purchase a lower level of NeuroSolutions, but then add in the Custom Solution Wizard to generate the neural networks as compiled DLLs instead of C++ source code.
Topologies
All topologies of the Users level
Hopfield networks
Time Delay Neural Network (TDNN)
Time-Lag Recurrent Network (TLRN)
General Recurrent
User-defined network topologies
Over 100 components to build from
A virtually infinite number of possible networks
Learning Paradigms
All paradigms of Users level
Recurrent backpropagation
Backpropagation through time
ANSI C++ source code generation
Embed neural networks into your own applications
Train networks on faster computers
Competitive Advantage
Unlimited inputs/outputs
Unlimited neurons per layer
Dynamical systems modeling
Time-Lag Recurrent Networks
Level C - Developers
The Developers level allows you to implement your own algorithms as DLLs and generate C++ source code for your neural networks. The generated code can even include your custom algorithms, giving you complete flexibility in the neural network development process.
Features
All features of the Professional level
User-defined dynamic link libraries to create customized neural components:
Nonlinearities
Interconnection matrices
Gradient search procedures
Error criteria
Unsupervised learning rules
Memory structures
Customized input/output
Customized parameter scheduling
MY Office : 72-3C, JALAN PUTERI 2/4, BANDAR PUTERI, 47100 PUCHONG, SELANGOR, Malaysia. Tel:+603-8063 9300 fax:+603-8063 9400
SG Office : 259, Onan Road, Singapore 424651. Tel: +65-6468 3325 Fax: +65-6764 5646