Neural Designer

The professional application for predictive analytics

Neural Designer is an advanced tool which makes intelligent use of data by discovering complex relationships, recognizing unknown patterns or predicting actual trends. The program uses neural networks, which are mathematical models of the brain function which can be trained to perform certain tasks.

Best results

When you're looking for a data mining software, the main idea is to perform specific business functions. Most predictive analytics tools contain a collection of different machine learning algorithms (neural networks, support vector machines, decision trees ...). Even specific neural networks software usually includes many different neural network models (multilayer perceptrons, radial basis networks, recurrent networks ...).

However, some algorithms are of purely academic interest, many of them present serious limitations, and several algorithms are equivalent to each other. Also, software tools that try to cover every method tend to lose focus, and often provide a very simple implementation for each.

Neural networks are the standard method for predictive analytics and are considered to be the best solution here. Our software is the most innovative implementation of that technique; it includes the results from many years of research resulting in a tool full of innovations. Don’t waste your time with packages that can't provide you the best information.

Easy to use

Usability is a measure of how easy it is to use a product to perform prescribed tasks. The interface characteristics are one of the main factors affecting the usability. However, many software tools do not provide an interface and must be operated by means of scripts (Python, R, MATLAB, ...). Here the users need to program their own applications. On the other hand, some user interfaces are unattractive, inconsistent and hard to interpret, such as block diagrams, technic that can be useful in electronic or control systems problems but for predictive analytics has no sense because problems are complex and not generics.

The training and support factors are also decisive. In this regard, undocumented or insufficiently documented software packages on data mining, machine learning and neural networks are quite common. Sometimes, the software developers do not offer training courses, or even technical support.

Our user interface has been conceived for simplifying data entry and it's not necessary to program to solve a problem, as well, the program includes Neural Assistant, a tool for presenting documentation which has a set of tutorials about neural networks and the use of the software. It provides detailed and comprehensive results.

High performance

Training neural networks with big data set can be very expensive in computational terms. Many data mining, machine learning and neural networks software are programmed using interpreted languages (Python, R, MATLAB ...) or semi-interpreted languages (Java). Interpreted languages have some advantages, but their drawback is the performance. In terms of time, they provide much slower program executions when compared to compiled languages. Regarding resources, they require more memory than compiled languages.

Neural designer outstands in performance, the amount of useful work accomplished by a system compared to the time and resources used. This has been achieved because Neural Designer has been created using the C++ programming language, which provides outstanding results in terms of execution speed and memory allocation.

As well, the program has been subject to optimization techniques to maximize the efficiency.