Metabolomics dataset from 165 different plasma measurements from 94 patients (cases) with glioma (brain tumours) and 71 healthy controls. The goal is to predict whether a sample is from a patient or a control based on the metabolite measurements.
glioma165 observations from 138 variables represented as a list consisting
of a binary response (factor) vector y with levels 'control' and 'case'
indicating whether the sample is from a healthy control or a patient with
glioma, as well as x: a matrix of 138 metabolite measurements.
We have removed the patients with meningioma from the original dataset (which contained 235 samples) to create a binary classification problem. Also, the authors originally had 188 features but removed some of these due to missing data.