Explore approximation by artificial neural networks using this intuitive GUI based on IGLib, which facilitates modeling and visualization.
- NeurApp
- Version :1.3
- License :Trial
- OS :Windows All
- Publisher :Igor Grešovnik, Tadej Kodelja
NeurApp Description
NeurApp (Neural Approximate) is a user-friendly GUI based on IGLib, which can help you explore approximation by artificial neural networks (ANNs). It supports both 1D and 2D artificial neural network models, giving you the possibility to specify parameters and plot results.
Explore approximation by ANNs
It’s not necessary to setup this tool since you can unzip the downloaded archive and double-click the .exe to reach the main app window. However, it cannot work unless you have .NET Framework or Mono installed on your computer.
The interface is based on a large window with two tabs for separately configuring settings when it comes to 1D and 2D approximation. For 1D mode, you can enable a function defined by the user and set the number of training samples, along with bounds.
Set properties for 1D and 2D approximation models
For 2D approximation mode, it’s possible to specify the training points on the X and Y axis, as well as to customize visualization settings related to the training points, original, approximation and contour graphs.
In both cases, you can also set application error values (maximum and RMS training and verification), enter the number of neurons in hidden layers, set the max epochs and epochs in bundle, enter the RMS, learning rate and momentum, as well as indicate the input and output safety factor.
The network can be trained with one click when everything is ready in NeurApp. Also, you can reset everything to default to restart from scratch. Unfortunately, there are no options implemented for copying the graph or data to the clipboard, exporting them to files, or printing them.
Easy-to-use artificial neural network explorer
The tool worked smoothly on Windows 10 in our tests. It had minimal impact on the computer’s performance and generated neural network models quickly.
All aspects considered, NeurApp offers a simple solution for producing ANN models based on 1D and 2D approximation settings.