Nelson-Aalen FeatureCloud App

Description

This app calculates the hazard rate for time-to-event data using the Nelson-Aalen estimator.

Input

input: containing the local training data (columns: features; rows: samples)

Output

cum_hazard_function: CSV file containing the survival function data
hazard_plot: PNG image showing the survival plot (Kaplan-Meier plot)
logrank_test: CSV file containing the pairwise logrank-test results

Workflows

This app is not compatible with other FeatureCloud apps.

Config

Use the config file to customize your training. Just upload it together with your training data as config.yml

fc_nelson_aalen:
  files: # file names
    input: lymphoma1.csv # name of the input CSV/TSV/sas7bdat file
    output: # name of the output files
      cum_hazard_function: "cum_hazard_function" # name of the CSV file containing the survival function data
      hazard_plot: "hazard_plot" # name of the PNG image showing the survival plot (Kaplan-Meier plot)
      logrank_test: logrank_test # If a category column is given: name of CSV file containing the pairwise logrank-test results

  # parameters
  parameters:
    duration_col: "Time" # name of the column containing the time values
    event_col: "Censor" # name of the column containing the event values (1=event occurred, 0=censored)
    category_col: "Stage_group" # name of the column containing the categories that shall be analysed separately (e.g. treatment A vs. treatment B)
    differential_privacy: "none" # amount of differential privacy added to the computation (none, low, middle or high)
    multipletesting_method: "bonferroni" # Method used for testing and adjustment of pvalues in the pairwise logrank test
      #bonferroni : one-step correction
      #sidak : one-step correction
      #holm-sidak : step down method using Sidak adjustments
      #holm : step-down method using Bonferroni adjustments
      #simes-hochberg : step-up method (independent)
      #hommel : closed method based on Simes tests (non-negative)
      #fdr_bh : Benjamini/Hochberg (non-negative)
      #fdr_by : Benjamini/Yekutieli (negative)
      #fdr_tsbh : two stage fdr correction (non-negative)
      #fdr_tsbky : two stage fdr correction (non-negative)

Privacy

Event times and counts are exchanged
Differential privacy can be applied to make the resulting hazard rates private.

julianspaeth / fc-nelson-aalen-estimator