In this page, the Data Analysis approach of NOESYS software is demonstrated. Following a vast amount of studies, we tried to develop a generic software environment, capable to analyze any database, composed by continuous variables. The analysis process follows four steps:

  1. Multiple correlations among all the variables and representation on a clustering map.
  2. Higher-order, polynomial regression combined with an optimization algorithm to automatically reveal the most significant non-linear features.
  3. Deep Neural Networks, trained with noesys-opti, for high generalization performance.
  4. Sensitivity analysis with Regression, ANN, and nearest neighborhoods models to thoroughly investigate features importance.
previous arrow
next arrow