This script trains and evaluates machine learning models for binary classification tasks. It supports multiple classification models, including Random Forest, KNN, SVM, Logistic Regression, and XGBoost.

• Train and evaluate multiple models for binary classification.
• Plot ROC curves and calculate AUC scores for model comparison.
• Calculate and plot feature importance using XGBoost models.
• Generate SHAP summary plots for model interpretability.

To run this script, you will need the following packages:

• pandas
• scikit-learn
• xgboost
• matplotlib
• joblib
• shap
• numpy

You can install these packages using the following command:

pip install pandas scikit-learn xgboost matplotlib joblib shap numpy

To use this script, you need to provide a CSV data file and specify the feature columns. The data file should contain a column named "Label" for the target labels and other columns for the features.

Example command:

python binary_classification.py data.csv 1-5

In this example, data.csv is the path to the CSV data file, and 1-5 specifies that columns 1 to 5 (inclusive) are the feature columns.

• data_file: Path to the CSV data file.
• feature_columns: Feature column range (e.g., 1,3,6 or 1-5 or 3).

• Models trained for each class pair are saved in the models directory.
• ROC curves are saved in the plots directory.
• Feature importance plots are saved in the plots directory.
• SHAP summary plots are saved in the plots directory.
• A CSV file named model_performance_results.csv containing the performance metrics of the trained models.

The main components of the script are:

• create_binary_labels(): Creates binary labels for a given class pair.
• split_data(): Splits the data into training and testing sets.
• plot_auc_combined(): Plots the ROC curves for all models.
• train_and_evaluate_all(): Trains and evaluates all models for a given class pair.
• get_feature_importance(): Retrieves the feature importance from an XGBoost model.
• plot_shap_values(): Plots the SHAP values for an XGBoost model.
• main(): The main function that orchestrates the training and evaluation process.

• The script assumes that the target column in the data file is named "Label".