Table 3 Example of input dataset showing a sample per row. Missing values have been represented by the NaN (Not a Number) literal. An excessive number of missing values will cause the elimination of the sample, while the remaining data will be statistically imputed in the preprocessing phase. Each sample is identified by its Id, and other relevant features are reported, e.g. the Age of the patient. The Disease feature holds the class information for each tuple (sample). The features on the right of the Disease column are protein expression values related to each sample

From: Machine learning pipeline to analyze clinical and proteomics data: experiences on a prostate cancer case