Controlled capacity modeling tool

A process for modeling numerical data for forecasting a phenomenon from a data set relating to the phenomenon, the process comprising:

• collecting data with a data acquisition module for describing the phenomenon;
• processing the data in a data preparation module to enhance exploitability of the data;
• constructing an initial model in a modeling module by learning on the processed data, the initial model having a fit to the data, robustness and parameters;
• evaluating the fit and robustness of the initial model in a performance analysis module;
• and
• adjusting the model parameters in an optimization module to select an optimal model, wherein the optimal model is generated in the form of a Dth order polynomial derived from variables input into the modeling module by controlling a trade-off between learning accuracy and learning stability, the trade-off being controlled by adding to a covariance matrix a perturbation in the form of a product of a scalar λ times a matrix H or in the form of a matrix H dependent on a vector of k parameters Λ=(λ1,λ2, . . . λk) during calculation of the optimal model, wherein the order D of the polynomial and the scalar λ, or the vector of parameters Λ, are determined automatically during model adjustment by the optimization module by integrating an additional data partition step performed by a partition module, the data partition step comprising the step of constructing a first subset comprising training data used as a learning base for the modeling module and a second subset comprising generalization data used to build a model validity criterion to adjust the value of the order D, the scalar λ or the vector of parameters Λ, and wherein the matrix H is a positive defined matrix of dimensions equal to a number p of variables input into the modeling module, plus one.