Use of High Dimensional Modeling for automatic variables selection: the best path algorithm

In the last decade the challenge of feature selection becomes a problems in different fields of the research [Li and Liu, 2017]. The dimensionality reduction is a strategy to solve this challenge. Although overtime scholars developed different methods, recently these methods have been challenged by Big Data Problem, in which the increasing availability of data is calling for new techniques able to handle not only a large amount of observations, but also rich data sets in terms of number and relations among variables [Yan et al., 2006]. In general, a dimensionality reduction problem can be viewed as an optimization problem, over a matrix of data: X [Saxena and Deb, 2008].