%store -r usa_binary
rfr = RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
            max_depth=5, max_features=0.2, max_leaf_nodes=None,
            min_impurity_decrease=0.0, min_impurity_split=None,
            min_samples_leaf=3, min_samples_split=2,
            min_weight_fraction_leaf=0.0, n_estimators=138, n_jobs=1,
            oob_score=False,
            verbose=0, warm_start=False)

accuracy = []
selectedFeatures = []
precs = []
for i in range(0, len(usa_binary), 7):
    print(i)
    train = fromListToSets([usa_binary[i]], 1, removeI = True)
    accuracyMax = 0
    features = []
    prec = 10
    if (train.shape[0] > 40):
        for i in [10, 20, 30, 40, 50, 60, 70]:
            try:
                feat_selector = BorutaPy(rfr, n_estimators='auto', random_state=1, perc = i, verbose=0, alpha=0.05)
                feat_selector.fit(train[:, 1:], train[:, 0])
                X_filtered = feat_selector.transform(train[:, 1:])
                k = 20
                scores = np.zeros((k))
                kFold = KFold(n_splits=k, random_state=None)
                y_pred = []
                if X_filtered.shape[1] == 0:
                    continue
                for train_index, test_index in kFold.split(train[:, 0]):
                    train_X = X_filtered[train_index, :]
                    train_Y = train[:, 0][train_index]
                    test_X = X_filtered[test_index, :]
                    test_Y = train[:, 0][test_index]
                    rfr.fit(train_X, train_Y)
                    pred = rfr.predict(test_X)
                    y_pred.append(pred)
                y_pred = list(itertools.chain.from_iterable(y_pred))
                accuracyForTree = np.sum(y_pred == train[:, 0]) / len(train[:, 0])
                if accuracyMax < accuracyForTree:
                    accuracyMax = accuracyForTree
                    features = feat_selector.support_
                    prec = i
            except (KeyboardInterrupt, SystemExit):
                raise
            except:
                pass
        accuracy.append(accuracyMax)
        selectedFeatures.append(features)
        precs.append(prec)
    else:
        accuracy.append(0)
        selectedFeatures.append(0)
        precs.append(0)