Research on the updated (higher dimensional) steroidogenic data from GC/MS

In 2018 the GC/MS dataset from IMSR, Birmingham was updated along with our research approach to it. Thus instead of the filtered 165 ratios we ended up with 496 ratios (ratio of the 32 extracted hormone to the rest of the 31 others). The dataset still contained highly imbalanced classes, heterogeneous measurements, and missing values. The other difference in this dataset and that one I used for my submission to ESANN 2017 is that this updated dataset presented a 7 class problem instead of the initial 4 class problem.