Patient Specific Survival Prediction - File Format

Data files must be in Comma-Seperated Value (CSV) format. Each line in the file describes one sample in your data set. The first column should be the survival time (or other variable to predict). The second column should be a binary value describing whether this data is censored (0=uncensored, 1=censored; u=uncensored, c=censored). The remaining columns are the rest of the features for each sample. Columns can be numeric, or nominal (categorical).

Missing values are represented by blanks (for example, see the "Weight" feature for the first patient below).

Accepted Example Data Format:

  Label,Censored,Height,Weight,Eye_Color
  55,0,155,,Brown
  45,1,180,144,Blue
  96,0,,120,Brown
  22,0,144,190,Green
  ...etc...

  Label,Censored,Height,Weight,Eye_Color
  55,u,155,,Brown
  45,c,180,144,Blue
  96,u,,120,Brown
  22,u,144,190,Green
  ...etc...

Note: For data to make predictions on (i.e. test data), you can simply put a placeholder (e.g. 0) in the label & censored columns, since these columns are ignored.