Obtención de modelos matemáticos cerrados a partir de datos experimentales

Data from structured data bases may be treated such that closed (as opposed to heuristic) models are obtained. In order to do this several mathematical and/or computational tools are required:

  1. Multi-variate regression algorithms,
  2. Genetic Algorithms,
  3. Spline collocation
  4. Multi-layer Perceptron Networks.

The systematic application of these tools allows us to obtain a mathematical model which achieves machine learning without the need of unproved and cumbersome heuristics. Applications os these method have been applied (by the author) to characterize populations (such as the ones prone to Chicongunya); analyze customer groups (in a large international Bank); eveluate the behavior of upcoming products (in a large Mexican organization). In this talk we will discuss de above mentioned techniques and their application.