最近也在做这个，核心代码给你,如下：

import matplotlib.pyplot as plt

import numpy as np

from sklearn.datasets import make_classification

from sklearn.linear_model import LogisticRegression

from sklearn.ensemble import (RandomTreesEmbedding, RandomForestClassifier, GradientBoostingClassifier)

from sklearn.preprocessing import OneHotEncoder

from sklearn.model_selection import train_test_split

from sklearn.metrics import roc_curve,auc

from sklearn.externals import joblib

from sklearn.externals.six import StringIO

from sklearn import tree

import pydotplus

n_estimator = 4 # the number of base trees

X, y = make_classification(n_samples=100,n_features=4)

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3,random_state=0)

X_train, X_train_lr, y_train, y_train_lr = train_test_split(X_train,y_train,test_size=0.3)

grd = GradientBoostingClassifier(n_estimators=n_estimator,max_depth=3)

grd.fit(X_train, y_train)

score_feature = grd.feature_importances_

print(score_feature)

joblib.dump(grd,'gbdt_model/grd_model.m',compress = 3)

m1 = grd.apply(X_train)[:, :, 0]

# print (m1)

grd_enc = OneHotEncoder()

grd_lm = LogisticRegression()

grd_enc.fit(m1)

joblib.dump(grd_enc,'gbdt_model/grd_enc_model.m',compress = 3)

grd_lm.fit(grd_enc.transform(grd.apply(X_train_lr)[:, :, 0]), y_train_lr)

joblib.dump(grd_lm,'gbdt_model/grd_lm_model.m',compress = 3)

# save the training model

m2 = grd_enc.transform(grd.apply(X_train_lr)[:, :, 0])

print (m2.toarray())

# grd = joblib.load('gbdt_model/grd_model.m')

# grd_enc = joblib.load('gbdt_model/grd_enc_model.m')

# grd_lm = joblib.load('gbdt_model/grd_lm_model.m')

y_pred_grd_lm = grd_lm.predict_proba(grd_enc.transform(grd.apply(X_test)[:, :, 0]))

print grd_lm.coef_

#print y_pred_grd_lm

acc = grd_lm.score(grd_enc.transform(grd.apply(X_test)[:, :, 0]),y_test)

print "acc is ",acc

fpr_grd_lm, tpr_grd_lm, _ = roc_curve(y_test, y_pred_grd_lm[:,1])

roc_auc = auc(fpr_grd_lm,tpr_grd_lm)

print "auc is ",roc_auc

dot_data = StringIO()

tree.export_graphviz(grd.estimators_[0,0],out_file = dot_data,node_ids=True,filled=True,rounded=True,special_characters=True)

graph = pydotplus.graph_from_dot_data(dot_data.getvalue())

graph.write_pdf("yang.pdf")

print('Visible tree plot saved as pdf.')