Multivariate analysis is the branch of statistics that generalizes methods of inferential statistics, so that a population can be characterized through a finite collection of random explicatives variables. As in classical inferential statistics, multivariate analysis the main idea is to generalize parameters or obtain useful conclusions from a multivariate population based on the information of the sample however in this case the information is multidimensional.
There are often situations in which it is necessary to make inferences about the future behavior of one or several variables in terms of random vectors, infer the population type of a random vector since there are several populations that share the same explicative variables but with different distribution, or find boundaries and structures of clustering since there are different types of mixed populations of which the membership of the vectors is not known. For these situations and some more, exist results based on multivariate analysis that provides methods for a non-exactly teoric solution to the problem, this is called statistical learning. In addition to this, a computational approach is added considering algorithms, complexity, expenditure, data structures, etc. then the set of these techniques known as machine learning.
