Feature selection in small-sample high-dimensional datasets enhances classification accuracy and reduces computational time for model training. This paper introduces the filter Classification Error Impurity (CEI) as a ...