clustering.py 文件源码

def cluster_gaussian(X_train, model_args=None, gridsearch=True):
    from sklearn.mixture import GaussianMixture
    print('GaussianMixture')

    if gridsearch is True:
        param_grid = {
            'n_components': np.arange(1, 20, 4),
            'covariance_type': ['full', 'tied', 'diag', 'spherical'],
        }
        prune(param_grid, model_args)
    else:
        param_grid = None

    return ModelWrapper(GaussianMixture, X=X_train, model_args=model_args, param_grid=param_grid, unsupervised=True)