-
Notifications
You must be signed in to change notification settings - Fork 0
/
2.3亚类划分-敏感度分析.py
179 lines (157 loc) · 6.46 KB
/
2.3亚类划分-敏感度分析.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
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
"""根据选定的最佳k值训练kmeans模型，赋给每个样本一个亚类标签，并以此训练决策树模型，得到具体的亚类划分方法，并对训练后的决策树模型进行敏感度测试"""
import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans
from sklearn.tree import DecisionTreeClassifier, plot_tree
import plotly.graph_objects as go
from PIL import Image
from pyecharts import options as opts
from pyecharts.charts import Line
from pyecharts.globals import ThemeType
import warnings
warnings.filterwarnings('ignore')
def plot_line_accu(y, y_label, title):
x = [str(i) + "%" for i in range(1, 41)]
trace = go.Scatter(
x=x,
y=y,
mode='lines',
)
layout = go.Layout(
xaxis=dict(
title='随机波动幅度',
ticks='outside',
),
yaxis=dict(
title=y_label,
ticks='outside',
),
)
fig = go.Figure(data=trace, layout=layout)
fig.show()
filename = "image/决策树" + title + y_label + ".png"
fig.write_image(filename, scale=6)
Image.open(filename).show()
def plot_line_prob(y1, y2, y_label1, y_label2, title):
c = (
Line(init_opts=opts.InitOpts(theme=ThemeType.MACARONS))
.add_xaxis(xaxis_data=[str(i) + "%" for i in range(1, 41)])
.add_yaxis(
series_name=y_label1,
y_axis=y1,
label_opts=opts.LabelOpts(is_show=False),
is_smooth=True
)
.add_yaxis(
series_name=y_label2,
y_axis=y2,
label_opts=opts.LabelOpts(is_show=False),
is_smooth=True
)
.set_global_opts(
xaxis_opts=opts.AxisOpts(name='随机波动幅度'),
yaxis_opts=opts.AxisOpts(
name="预测概率值",
type_="value",
splitline_opts=opts.SplitLineOpts(is_show=True),
min_=0,
max_=1
),
toolbox_opts=opts.ToolboxOpts(
feature=opts.ToolBoxFeatureOpts(
save_as_image=opts.ToolBoxFeatureSaveAsImageOpts(
pixel_ratio=6, # 设置保存图片的分辨率
background_color='white', # 设置导出图片的背景颜色
)
)
),
)
.render('output/决策树' + title + '预测概率值.html')
)
def sens_analyse(model, X, y_true, y_label1, y_label2, title):
scores = []
pred_probas1 = []
pred_probas2 = []
for _ in range(1000):
score_round = []
proba_round1 = []
proba_round2 = []
for i in range(1, 41):
X_wave = X * (np.random.uniform(low=(100 - i) / 100, high=(100 + i) / 100, size=(len(X), 14)))
score_round.append(model.score(X_wave, y_true))
proba_round1.append(model.predict_proba(X_wave[[3], :]))
proba_round2.append(model.predict_proba(X_wave[[7], :]))
scores.append(score_round)
pred_probas1.append(proba_round1)
pred_probas2.append(proba_round2)
scores = np.array(scores).mean(axis=0)
scores = [round(score, 2) for score in scores]
pred_probas1 = np.array(pred_probas1).mean(axis=0)
pred_probas2 = np.array(pred_probas2).mean(axis=0)
pred_probas1 = [round(max(proba[0]), 2) for proba in pred_probas1]
pred_probas2 = [round(max(proba[0]), 2) for proba in pred_probas2]
print(scores)
plot_line_accu(scores, '精确度', title)
plot_line_prob(pred_probas1, pred_probas2, y_label1, y_label2, title)
pd.set_option('display.max_columns', None)
pd.set_option('display.max_rows', None)
plt.rcParams['font.sans-serif'] = ['SimHei'] # 正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False # 正常显示负号
plt.style.use('ggplot')
filename = 'data/附件.xlsx'
info = pd.read_excel(filename, sheet_name='表单1')
chem = pd.read_excel(filename, sheet_name='表单2')
chem.fillna(0, inplace=True)
chem['累加和'] = chem.iloc[:, 1:].apply(lambda x: x.sum(), axis=1)
chem = chem[chem['累加和'] <= 105]
chem = chem[chem['累加和'] >= 85]
chem['文物编号'] = chem['文物采样点'].apply(lambda x: int(x[:2]))
chem['采样点'] = chem['文物采样点'].apply(lambda x: x[2:])
chem.index = list(range(len(chem)))
chem['类型'] = [0] * len(chem)
chem['表面风化'] = [0] * len(chem)
for i in chem.index:
chem['类型'][i] = info[info['文物编号'] == chem['文物编号'][i]]['类型'].tolist()[0]
chem['表面风化'][i] = info[info['文物编号'] == chem['文物编号'][i]]['表面风化'].tolist()[0]
highK = chem[chem['类型'] == '高钾']
highK_X = highK.iloc[:, list(range(1, 15))]
highK_X = StandardScaler().fit_transform(highK_X)
kmeans = KMeans(n_clusters=6)
kmeans.fit(highK_X)
highK['亚类'] = kmeans.labels_
print(highK.loc[:, ['二氧化硅(SiO2)', '二氧化硫(SO2)', '氧化钙(CaO)', '氧化钡(BaO)', '氧化镁(MgO)', '亚类']])
highK_X = highK.iloc[:, list(range(1, 15))]
highK_y = highK['亚类']
# 初始化决策树分类器
dtc = DecisionTreeClassifier(random_state=0)
dtc.fit(highK_X, highK_y)
plot_tree(dtc,
feature_names=highK_X.columns,
class_names=['亚类Ⅰ', '亚类Ⅱ', '亚类Ⅲ', '亚类Ⅳ', '亚类Ⅴ', '亚类Ⅵ'],
filled=True,
rounded=True)
plt.savefig('image/高钾玻璃决策树.png', dpi=600)
highK_X = highK_X.to_numpy()
sens_analyse(dtc, highK_X, highK['亚类'], '文物编号4', '文物编号7', "高钾玻璃")
PbBa = chem[chem['类型'] == '铅钡']
PbBa_X = PbBa.iloc[:, list(range(1, 15))]
PbBa_X = StandardScaler().fit_transform(PbBa_X)
kmeans = KMeans(n_clusters=6)
kmeans.fit(PbBa_X)
PbBa['亚类'] = kmeans.labels_
PbBa_X = PbBa.iloc[:, list(range(1, 15))]
PbBa_y = PbBa['亚类']
# 初始化决策树分类器
dtc = DecisionTreeClassifier(random_state=0)
dtc.fit(PbBa_X, PbBa_y)
plot_tree(dtc,
feature_names=PbBa_X.columns,
class_names=['亚类Ⅰ', '亚类Ⅱ', '亚类Ⅲ', '亚类Ⅳ', '亚类Ⅴ', '亚类Ⅵ'],
filled=True,
rounded=True
)
plt.savefig('image/铅钡玻璃决策树.png', dpi=600)
PbBa_X = PbBa_X.to_numpy()
sens_analyse(dtc, PbBa_X, PbBa['亚类'], '文物编号11', '文物编号24', "铅钡玻璃")