More than one estimator to select Features from Model

[bjkal]

I created an own config_dict and my aim was to see which Estimators are choosed by TPOT to select Features from a Model.

Its working, if i choose only one estimator with the corresponding model parameters like that:

'sklearn.feature_selection.SelectFromModel': {
'threshold': np.arange(0, 1.01, 0.05),
'estimator': {
'sklearn.ensemble.ExtraTreesRegressor': {
'n_estimators': [100],
'max_features': np.arange(0.05, 1.01, 0.05)
}

The Problem occurs if i want TPOT to choose between different estimators like that:

'sklearn.feature_selection.SelectFromModel': {
'threshold': np.arange(0, 1.01, 0.05),
'estimator': {
'sklearn.ensemble.ExtraTreesRegressor': {
'n_estimators': [100],
'max_features': np.arange(0.05, 1.01, 0.05)
},
'sklearn.ensemble.RandomForestRegressor':{
},
'sklearn.linear_model.Ridge':{
},
'sklearn.linear_model.LassoCV':{
}
}}

How i have to create/change the dictionary config_dict ?