-
Notifications
You must be signed in to change notification settings - Fork 0
/
tree_extreme_plots.py
113 lines (86 loc) · 3.29 KB
/
tree_extreme_plots.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
#tree_extreme_plots.py
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import xgboost as xgb
import time
import pdb
def plotty_plots(diff, model, parm_file):
diff_train, diff_test = diff
fig2 = plt.figure()
plt.scatter(range(0,np.shape(diff_train)[0]),diff_train, s=1)
plt.title('RMSE error on train set')
plt.xlabel(' y- y_pred Error')
plt.ylabel('prediction error')
filename = "sparse_weight_" + time.strftime("%m%d_%H%M") + ".pkl"
filename = './output/xgb_rmse_train_' + time.strftime("%m%d_%H%M") + '.png'
plt.savefig(filename)
# plt.show()
# plt.close()
fig3 = plt.figure()
plt.scatter(range(0,np.shape(diff_test)[0]),diff_test, s=1)
plt.title('RMSE error on test set')
plt.xlabel(' y- y_pred Error')
plt.ylabel('prediction error')
filename = './output/xgb_rmse_test_' + time.strftime("%m%d_%H%M") + '.png'
plt.savefig(filename)
#### tree plots -- will slow down your system if max depth is too big
# pdb.set_trace()
# fig1,ax = plt.figure()
# plt.figure(2)
# xgb.plot_tree(model, num_trees=0,rankdir='LR')
#
# plt.figure(3)
# xgb.plot_tree(model, num_trees=1,rankdir='LR')
# plt.figure(4)
# xgb.plot_tree(model, num_trees=2,rankdir='LR')
# plt.figure(5)
# xgb.plot_tree(model, num_trees=3,rankdir='LR')
# plt.figure(6)
fig, ax = plt.subplots(figsize=(100, 100))
ax = xgb.plot_tree(model, num_trees=5,rankdir='LR', ax = ax)
filename = './output/xgb_tree_plot' + time.strftime("%m%d_%H%M") + '.png'
plt.savefig(filename)
# plt.show()
fig4 = plt.figure()
sns.set(color_codes=True)
ls_diff = list(diff_train)
ax = sns.distplot(ls_diff, kde=False, rug=True)
ax.set_title('RMSE error distribution on train set')
filename = './output/xgb_rmse_distribution_train_' + time.strftime("%m%d_%H%M") + '.png'
plt.savefig(filename)
fig5 = plt.figure()
sns.set(color_codes=True)
ls_diff = list(diff_test)
ax = sns.distplot(ls_diff, kde=False, rug=True)
ax.set_title('RMSE error distribution on test set')
filename = './output/xgb_rmse_distribution_test_' + time.strftime("%m%d_%H%M") + '.png'
plt.savefig(filename)
# plt.show()
fig6 = plt.figure()
xgb.plot_importance(model, max_num_features=10)
# plt.rcParams['figure.figsize'] = [15, 15]
filename = './output/xgb_importance_plot' + time.strftime("%m%d_%H%M") + '.png'
plt.savefig(filename)
# plt.show()
fig7 = plt.figure()
feature_important = model.get_score(importance_type='weight')
keys = list(feature_important.keys())
values = list(feature_important.values())
importance = pd.DataFrame(data=values, index=keys, columns=["score"])
# pdb.set_trace()
importance.nlargest(20, columns='score').plot(kind='barh')
most_important = list(importance.nlargest(20, columns='score').index)
filename = './output/most_important' + time.strftime("%m%d_%H%M") + '.txt'
f=open(filename, "w")
# f.write('\n')
# f.write('top 50 features:: '+'\n')
for ff in most_important:
f.write(ff)
f.write('\n')
f.close()
filename = './output/xgb_importance_plot_by score' + time.strftime("%m%d_%H%M") + '.png'
plt.savefig(filename)
# plt.show()
# pdb.set_trace()