classifierpromax.ClassifierOptimizer

Functions

get_param_distributions()
get_scoring_metrics()
validate_inputs(model_dict, param_dist, X_train, ...)
optimize_model(name, model, param_dist, X_train, ...)
evaluate_model(name, model, X_train, y_train, cv)
ClassifierOptimizer(model_dict, X_train, y_train[, ...])	Optimizes a dictionary of scikit-learn Pipeline classifiers using RandomizedSearchCV

Module Contents

classifierpromax.ClassifierOptimizer.get_param_distributions()

classifierpromax.ClassifierOptimizer.get_scoring_metrics()

classifierpromax.ClassifierOptimizer.validate_inputs(model_dict, param_dist, X_train, y_train, scoring, n_iter, cv, random_state, n_jobs)

classifierpromax.ClassifierOptimizer.optimize_model(name, model, param_dist, X_train, y_train, scoring, n_iter, cv, random_state, n_jobs)

classifierpromax.ClassifierOptimizer.evaluate_model(name, model, X_train, y_train, cv)

classifierpromax.ClassifierOptimizer.ClassifierOptimizer(model_dict, X_train, y_train, scoring='f1', n_iter=100, cv=5, random_state=42, n_jobs=-1)

Optimizes a dictionary of scikit-learn Pipeline classifiers using RandomizedSearchCV and evaluates their performance.

Parameters:

model_dictdict

A dictionary where keys are model names (str) and values are scikit-learn Pipeline objects. Each pipeline must contain a classifier whose hyperparameters are defined in param_dist.

X_trainpandas.DataFrame or numpy.ndarray

The feature matrix for training the classifiers. Must have the same number of samples as y_train.

y_trainpandas.Series or numpy.ndarray

The target labels for training the classifiers. Must have the same number of samples as X_train.

scoringstr, optional, default=’f1’

The scoring metric to use for hyperparameter optimization and model evaluation. Must be one of the following: - “accuracy” - “precision” - “recall” - “f1”

n_iterint, optional, default=100

The number of parameter settings sampled for RandomizedSearchCV.

cvint, optional, default=5

The number of cross-validation folds for both RandomizedSearchCV and cross_validate.

random_stateint, optional, default=42

Random seed for reproducibility of RandomizedSearchCV.

n_jobsint, optional, default=-1

The number of jobs to run in parallel for RandomizedSearchCV (-1 uses all available processors).

Returns:

optimized_model_dictdict

A dictionary containing the best estimators for each classifier after hyperparameter optimization.

scoring_dictdict

A dictionary containing cross-validation results for each optimized model, with metrics aggregated by mean and standard deviation.

Raises:

ValueError

If the input parameters are invalid (e.g., empty model dictionary, mismatched data shapes, unsupported scoring metric).

Examples:

>>> from sklearn.pipeline import Pipeline
>>> from sklearn.linear_model import LogisticRegression
>>> from sklearn.svm import SVC
>>> from sklearn.ensemble import RandomForestClassifier
>>> from sklearn.preprocessing import StandardScaler
>>> model_dict = {
...     'logreg': Pipeline([
...         ('scaler', StandardScaler()),
...         ('logisticregression', LogisticRegression())
...     ]),
...     'svc': Pipeline([
...         ('scaler', StandardScaler()),
...         ('svc', SVC())
...     ]),
...     'random_forest': Pipeline([
...         ('randomforestclassifier', RandomForestClassifier())
...     ])
... }
>>> optimized_models, scoring_results = ClassifierOptimizer(model_dict, X_train, y_train)