Classification by ensembles from random partitions of high-dimensional data

摘要: A robust classification procedure is developed based on ensembles of classifiers, with each classifier constructed from a different set predictors determined by random partition the entire predictors. The proposed methods combine results multiple classifiers to achieve substantially improved prediction compared optimal single classifier. This approach designed specifically for high-dimensional data sets which sought. By combining built subspace predictors, computational advantage in tackling growing problem dimensionality. For we build tree or logistic regression tree. Our study shows, using four real areas, that our perform consistently well widely used methods. unbalanced data, maintains balance between sensitivity and specificity more adequately than many other considered this study.