Statistical analysis methods with scattered data Métodos de análisis estadístico con datos dispersos

Abstract

This article presents the main characteristics of some methodologies to derive prediction or classification models when you have a lot of data. These describe neural networks, random forests, decision trees, highly randomized trees (extra trees) and the contrast pattern assisted regression method (CPXR). Techniques that have been used to derive prediction models for water retention, image classification, soil classification, determination of the main variables involved in a process.