Module special_model_predictions

Function leave_out_prediction

View Source

leave_out_prediction(model: LogisticRegression | RandomForestClassifier | Pipeline, features: DataFrame, gt: Series, patient_ids: Series) -> DataFrame

Trains model on all data except one sample and predicts probability on the left out sample. Results are stored in a DataFrame. Args: model (Union[LogisticRegression, RandomForestClassifier, Pipeline]): Model to train features (pd.DataFrame): Features of model gt (pd.Series): Groundtruth labels

---