>>> from sklearn.datasets import load_iris
>>> from sklearn.feature_selection import SelectKBest
>>> from sklearn.feature_selection import chi2
>>> iris = load_iris()
>>> X, y = iris.data, iris.target
>>> X.shape
(150, 4)
>>> X_new = SelectKBest(chi2, k=2).fit_transform(X, y)

from statsmodels.stats.outliers_influence import variance_inflation_factor

def check_vif(data, test, target_name, cols_to_drop, verbose=0):
    if verbose == 1:
        print('Checking VIF values')
    X_train_multicoll = data.drop([target_name], axis=1).copy()
    X_train_multicoll['intercept'] = 1
    max_vif_value = float('inf')
    if verbose == 1:
        print(X_train_multicoll.shape)
    while max_vif_value > 100:
        vif = [variance_inflation_factor(X_train_multicoll.values, i) for i in
                range(X_train_multicoll.shape[1])]
        g = [i for i in list(zip(X_train_multicoll.columns, vif))]
        g = [i for i in g if i[0] != 'intercept']
        max_vif = max(g, key=itemgetter(1))
        if verbose == 1:
            print(max_vif)
        if max_vif[1] < 100:       
            if verbose == 1:
                print('Done')
            break
        else:
            X_train_multicoll.drop([max_vif[0]], axis=1, inplace=True)
            cols_to_drop.append(max_vif[0])
            data.drop([max_vif[0]], axis=1, inplace=True)
            test.drop([max_vif[0]], axis=1, inplace=True)
        if verbose == 1:
            print(X_train_multicoll.shape)
        max_vif_value = max_vif[1]
    return data, test, cols_to_drop

def permutation_importances(rf, X_train, y_train, metric):
    baseline = metric(rf, X_train, y_train)
    imp = []
    for col in X_train.columns:
        save = X_train[col].copy()
        X_train[col] = np.random.permutation(X_train[col])
        m = metric(rf, X_train, y_train)
        X_train[col] = save
        imp.append(baseline - m)
    return np.array(imp)

perm = PermutationImportance(model, random_state=1).fit(X_train, y_train)
eli5.show_weights(perm, top=50)

from boruta import BorutaPy
rf = RandomForestClassifier(n_jobs=-1, class_weight='balanced', max_depth=5)

# define Boruta feature selection method
feat_selector = BorutaPy(rf, n_estimators='auto', verbose=2, random_state=1)

# find all relevant features - 5 features should be selected
feat_selector.fit(X, y)

# check selected features - first 5 features are selected
feat_selector.support_

# check ranking of features
feat_selector.ranking_

# call transform() on X to filter it down to selected features
X_filtered = feat_selector.transform(X)

from boostaroota import BoostARoota

br = BoostARoota(metric='logloss')

#Fit the model for the subset of variables
br.fit(x, y)

#Can look at the important variables - will return a pandas series
br.keep_vars_

#Then modify dataframe to only include the important variables
x1 = br.transform(x)

features = X_train.columns
X_train['target'] = 0
X_valid['target'] = 1

train_test = pd.concat([X_train, X_valid], axis =0)

target = train_test['target']

# train model