Train Random Forest on Zarr Data with Cross-Validation

A solution that trains a Random Forest model using data from a Zarr zip store, filters runs with only one label, and performs cross-validation.

Arguments

--input_zarr_path

Path to the input Zarr zip store containing the features and labels. (default value: PARAMETER_VALUE)

--output_model_path

Path for the output joblib file containing the trained Random Forest model. (default value: PARAMETER_VALUE)

--n_estimators

Number of trees in the Random Forest. (default value: 200)

--max_depth

The maximum depth of the trees. (default value: 14)

--max_samples

The maximum number of samples to draw from X to train each base estimator. (default value: 0.495033308120533)

--min_samples_split

The minimum number of samples required to split an internal node. (default value: 2)

--min_samples_leaf

The minimum number of samples required to be at a leaf node. (default value: 1)

--class_weights

Class weights for the Random Forest model as a comma-separated list. (default value: )

--n_splits

Number of folds for cross-validation. (default value: 10)

Please follow this link for details on how to install and run this solution.