from sklearn.linear_model import SGDClassifier
from sklearn.datasets import load_iris
from sklearn import cross_validation

iris = load_iris()

hyperparameter_choices = [ 

    # some examples, by no means exhausive
    {u'loss': 'modified_huber', u'shuffle': True, u'n_iter': 25.0, 
    u'l1_ratio': 0.5, u'learning_rate': 'constant', u'fit_intercept': 0.0, 
    u'penalty': 'l2', u'alpha': 1000.0, u'eta0': 0.1, u'class_weight': None},

    {u'loss': 'squared_hinge', u'shuffle': True, u'n_iter': 25.0, u'l1_ratio': 0.5, 
    u'learning_rate': 'optimal', u'fit_intercept': 0.0, u'penalty': 'elasticnet', 
    u'alpha': 0.001, u'eta0': 0.1, u'class_weight': None},

    {u'loss': 'squared_hinge', u'shuffle': True, u'n_iter': 100.0, u'l1_ratio': 0.5, 
    u'learning_rate': 'optimal', u'fit_intercept': 0.0, u'penalty': 'l2', u'alpha': 0.001, 
    u'eta0': 0.001, u'class_weight': None}
]

for params in hyperparameter_choices:
    try:
        clf = SGDClassifier(**params)
        scores = cross_validation.cross_val_score(clf, iris.data, iris.target, cv=5)
    except ValueError as ve:
        print "ValueError: %s" % ve