Our approach is based on several Artificial Intelligence tools with superior performance over traditional statistical methods. They are:

Artificial Neural Networks (ANN),
ANN are non-parametric learning machines capable to discover complex, often hidden, patterns in the data. We used the classical MLP and a new variation invented by our team, the HLVQC, which is more accurate and robust.

Support Vector Machines (SVM)
SVM are a collection of state of the art algorithms to perform efficient highly complex tasks like classification of data and regression.

Multi-Objective Genetic Algorithms (MOGA)
MOGA are evolvable algorithms that, in most the same way as nature does, discover solutions for optimization problems. In our case is finding the best and smaller set of attributes to reach accuracy.

Isomap & NMF
Isomap is a modern non-linear algorithm that can be used to visualize complex data and reduce its complexity. NMF (non negative matrix factorization) is an algorithm do discover the relevant features hidden in complex data.

NOTE: AIRES is still in beta version. It will soon have implemented new features.