ML4HC_Interpretability_and_Explainatility

Part 1

Q1: Exploratory Data Analysis

• The p1/1_data_exploration.ipynb file contains all the necessary code for this section.
• It uses the raw data (train_val_split.csv and test_split.csv) in the p1/data folder and also saves processed data into the same folder (train_val.parquet and test.parquet).
• The file contains code for visualizing the feature distributions and processing the data.

Q2: Logistic Lasso Regression

• The code for the section is in the p1/2_lasso_regression.ipynb file.
• It contains all the necessary functions but uses the previously processed data in the p1/data folder (train_val.parquet, test.parquet).

note: The part of the code that saves the model and the results on the test set is commented out.

Q3: Multi-Layer Perceptrons

• The p1/3_multi-layer-perceptrons.ipynb file contains all the necessary functions and classes for this part.
• It also uses the processed data files in the p1/data folder (train_val.parquet, test.parquet).

note: The part of the code that saves the model and the results on the test set is commented out.

Q4: Neural Additive Models

• The same way as previously, the p1/4_neural_additive_models.ipynb file contains all the necessary functions and classes for this part.
• It also uses the processed data files in the p1/data folder (train_val.parquet, test.parquet).

note: The part of the code that saves the model and the results on the test set is, again, commented out.