crawles/automl_service

Parameters, data format?

Opened this issue · 1 comments

Is there any documentation on the data format or the parameters that you take in?

pipeline_id: 1
extract_features:
  column_id: example_id
  impute_function: tsfresh.utilities.dataframe_functions.impute
  default_fc_parameters: MinimalFCParameters()
model_training:
  model_args:
  model: sklearn.ensemble.RandomForestClassifier

The training specific parameters are specific to tsfresh, tpot, and sklearn.

  • extract_features
  • model - any sklearn-learn compatible classifier
  • model_args - arguments for associated sklearn classifier

For the data format, you have to use the same format as shown in the data folder. Check out the appendix of the Jupyter Notebook to see how I generated the data.