We have applied the Mitchell's criteria in the searching of new physics particles inside the framework of Supersymmetry considered to be candidate for a beyond Standard Model theory. On a pre-existent code that yields a low signal efficiency the implementation of the Machine Learning machinery have allows us to cut away background events in a 16% more than the case when traditional procedures for signal recovery are used. As result the Mitchell's strategy can improve the cut events procedure as seen in the final statistical error from a 3.70% to a 1.22%. Although a difference of 2.48% might be perceived as small it might be of relevance in those experiments where data can surpass 1T of events.