This paper introduces the application of feature ranking with a twofold purpose: first it achieves a feature sorting which becomes into a feature selection procedure given that a threshold is defined to only retain features with a positive influence with the class label (feature sorting and feature selection: FSoFS), second the feature subset is optimised using feature ranking to get a promising order of attributes (FSo). The supporting paper introduced a single stage: Feature ranking for feature sorting and feature selection, FR4(FS) with the shortcoming that for some high-dimensional classification problems the pre-processed data set did not obtain a more accurate classification model than the raw classifier. The results mean that it deserves to consider feature ranking in a couple of sequential stages with different goals: i) feature sorting, accompanied by feature selection, and ii) also alone