API-referentie

uitnodigingsregel

Uitnodigingsregel: dropout prediction models for student intervention.

detect_separator(file_path, target_column='Dropout')

Detect the CSV separator by trying common delimiters.

Reads the first 5 rows with each candidate separator and returns the first one that produces multiple columns containing the target column.

Parameters:

Name	Type	Description	Default
file_path	str | Path	Path to the CSV file.	required
target_column	str	Expected column name used to validate the separator.	'Dropout'

Returns:

Type	Description
str	Detected separator character, defaults to ',' if none matched.

load_settings(config_file=None, settings_type='default')

Load settings from YAML config file.

Parameters:

Name	Type	Description	Default
config_file	str | None	Path to config YAML. Defaults to package metadata config.	None
settings_type	str	Which settings block to load ('default' or 'custom').	'default'

Returns:

Type	Description
dict	Dictionary of settings values.

train_lasso(dataset_train_scaled, random_seed, dropout_column, alpha_range, model_path=Path('models/lasso_regression.joblib'))

Train a Lasso regression model with grid search over alpha values.

Parameters:

Name	Type	Description	Default
dataset_train_scaled	DataFrame	Scaled training DataFrame with features and target.	required
random_seed	int	Random state for reproducibility.	required
dropout_column	str	Name of the target column.	required
alpha_range	list[float]	List of alpha values to search.	required
model_path	Path	Path to save the trained model.	Path('models/lasso_regression.joblib')

Returns:

Type	Description
Lasso	Best-fit Lasso model.

train_random_forest(dataset_train, random_seed, dropout_column, rf_parameters, model_path=Path('models/random_forest_regressor.joblib'))

Train a Random Forest regressor with grid search hyperparameter tuning.

Parameters:

Name	Type	Description	Default
dataset_train	DataFrame	Training DataFrame with features and target.	required
random_seed	int	Random state for reproducibility.	required
dropout_column	str	Name of the target column.	required
rf_parameters	dict	Parameter grid for GridSearchCV.	required
model_path	Path	Path to save the trained model.	Path('models/random_forest_regressor.joblib')

Returns:

Type	Description
RandomForestRegressor	Best-fit RandomForestRegressor model.

train_svm(dataset_train_scaled, random_seed, dropout_column, svm_parameters, model_path=Path('models/support_vector_machine.joblib'))

Train an SVM classifier with grid search hyperparameter tuning.

Parameters:

Name	Type	Description	Default
dataset_train_scaled	DataFrame	Scaled training DataFrame with features and target.	required
random_seed	int	Random state for reproducibility.	required
dropout_column	str	Name of the target column.	required
svm_parameters	dict	Parameter grid for GridSearchCV.	required
model_path	Path	Path to save the trained model.	Path('models/support_vector_machine.joblib')

Returns:

Type	Description
SVC	Best-fit SVC model with probability estimates enabled.