Products & Solutions / By Solution / NeuroFusion / FAQ


FAQ


How does NeuroFusion get input data?
What about graphics?
What functions does the NeuroFusion have?
How can I evaluate NeuroFusion prior to purchase?
I need to customize your DLL. Can you help me?
Can I use your library to develop and globally distribute commercial software?
Do I have to pay royalties if I develop and ship commercial software using the library?
Is there any difference in price between NeuroFusion Smart and NeuroFusion Pro?


How does NeuroFusion get input data?
The library takes input data through a 2-dimensional matrix of input values. The library does not limit your flexibility with fixed input file formats. You can fill in the matrix however you wish. The matrix can contain floating points, integer and text (categorical) columns. In NeuroFusion Pro you can use the standard ,NET class DataTable to present your data.


What about graphics?
NeuroFusion is completely independent of graphic interfaces. The library itself does not provide any graphical output but you can use statistics provided by the library to draw training graphs, case errors graphs, actual vs. output graphs and scatter plots, ROC curves, response graphs, input importance charts and more.


What functions does the NeuroFusion have?
NeuroFusion Smart ensures easy integration of neural networks into your software. All theoretical details are hidden inside the library. To save your time a minimum function set of 15 functions is created. The basic set includes 9 functions for importing of input data, creating, deleting, training, applying, saving and restoring networks. The advanced functions are used for reporting on the network training progress and the current state, reinitializing/pausing/resuming/stopping network training, and for specifying the training conditions.

NeuroFusion Pro is object-oriented, so it provides several public classes that perform all required tasks. The main class is NetworkTrainer that trains new and existing neural models, provides various options and events. Resulting model is presented as a NeuralModel class: it can execute queries for new data, save and load itself.


How can I evaluate NeuroFusion prior to purchase?
We do not send evaluation versions but you can easily evaluate the NeuroFision Smart library performance using one of Alyuda products. The similar DLL is used in Alyuda Forecaster XL. You can comprehensively evaluate its speed and performance by testing Forecaster XL. You can download a free 30-day trial from Alyuda website download page.

NeuroFusion Pro uses the same network engine as Alyuda NeuroIntelligence that also can be downloaded for free.


I need to customize your DLL. Can you help me?
Yes. If you need to have the library customized we can do it on per hour basis. Send us your requirements and we will estimate the project time and costs free of charge.


Can I use your library to develop and globally distribute commercial software?
Yes, but you need to purchase the Integration License. See the licensing page for more details.


Do I have to pay royalties if I develop and ship commercial software using the library?
No. You do not have to pay per user charges; however, you will need to purchase the Integration License. See the licensing page for more details.


Is there any difference in price between NeuroFusion Smart and NeuroFusion Pro?
No. Each edition has its own advantages. The Pro Edition allows you to have a full control over the network’s architecture and training whereas the Smart Edition enables you to automate almost all operations.

top