chrisby/ml-on-fhir

Issues

• Provide pypi installation

  #37 opened 6 years ago by chrisby
  0

• Check import statements in examples and documentation after pypi

  #36 opened 6 years ago by chrisby
  0

• Proper preprocessor for case/controls in classification tasks.

  #26 opened 6 years ago by mtog
  0

• From a selected patient cohort generate a control cohort

  #12 opened 6 years ago by chrisby
  0

• Go through all modules and check & update documentation

  #25 opened 6 years ago by chrisby
  0

• Extend documentation to include ML

  #22 opened 6 years ago by chrisby
  0

• Use LOINC codes for observation preprocessing

  #17 opened 6 years ago by chrisby
  1

• Implement common observation preprocessors in preprocessor class.

  #19 opened 6 years ago by chrisby
  0

• Design preprocessing module as class

  #20 opened 6 years ago by chrisby
  0

• Finish automated registration of predefined ObservationPreprocessors

  #24 opened 6 years ago by chrisby
  0

• Visualize Clustering Results

  #2 opened 6 years ago by chrisby
  0

• Describe Preprocessor class in read the docs

  #23 opened 6 years ago by chrisby
  0

• Setup sphinx documentation

  #21 opened 6 years ago by chrisby
  0

• Add seed option when initializing MLOnFhir

  #14 opened 6 years ago by mtog
  0

• Simplify registration of observation processor

  #18 opened 6 years ago by chrisby
  0

• Report or store prediction performances

  #9 opened 6 years ago by chrisby
  0

• Adapt MLOnFhir class for clustering support

  #13 opened 6 years ago by mtog
  1

• More descriptive demo ipython notebook

  #11 opened 6 years ago by chrisby
  0

• Update Patient based on classification results

  #3 opened 6 years ago by chrisby
  0

• Allow to return multi dimensional features for a single attribute

  #16 opened 6 years ago by chrisby
  0

• Introduce imputation for missing features

  #15 opened 6 years ago by chrisby
  0

• Extend FHIR client to query subpopulations

  #6 opened 6 years ago by chrisby
  1

• Create a larger body of test patients for more interesting examples

  #10 opened 6 years ago by chrisby
  1

• Implement kNN clf on a single attribute of patients

  #1 opened 6 years ago by chrisby
  2

• Allow for preprocessors to register from outside.

  #7 opened 6 years ago by chrisby
  0

• Make a standard cross validation setup default for training.

  #8 opened 6 years ago by chrisby
  0

• Make preprocessing possible for multiple attributes

  #5 opened 6 years ago by chrisby
  0

• Remove ugly dict structure

  #4 opened 6 years ago by chrisby
  0