CellPhenoX Module

CellPhenoX Class

class pyCellPhenoX.CellPhenoX(X, y, CV_repeats, outer_num_splits, inner_num_splits)

Bases: object

__init__(X, y, CV_repeats, outer_num_splits, inner_num_splits)

_summary_

Parameters:

• X (dataframe) – cell by latent dimensions dataframe

• y (series) – the target variable

• CV_repeats (int) – number of times to repeat the cross-validation

• outer_num_splits (int) – number of outer folds (stratified k fold)

• inner_num_splits (int) – number of inner folds (for hyperparameter tuning)

get_best_model()

get_best_score()

get_interpretable_score()

get_model_training_time()

get_prc_curves()

get_roc_curves()

get_shap_values(outpath)

get_shap_values_per_cv()

model_training_shap_val(fast, outpath)

Train the model using nested cross validation strategy and generate shap values for each fold/CV repeat

Parameters: fast (bool): whether to use the fasttreeshap package or the shap package outpath (str): the path for the output folder

Returns:

shap_values_explainer(model, X, fast=True, n_jobs=-1)

split_data(train_outer_ix, test_outer_ix)