/
forest_cover_type.py
executable file
·249 lines (192 loc) · 8.61 KB
/
forest_cover_type.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
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
#!/usr/bin/env python2.7
# -*- coding: utf-8 -*-
from pandas import read_csv, DataFrame, Series, concat
from sklearn.preprocessing import LabelEncoder
from sklearn import cross_validation, svm, grid_search
from sklearn.neighbors import KNeighborsClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import ExtraTreesClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.metrics import roc_curve, auc
from sklearn.metrics import accuracy_score, roc_auc_score
import pylab as pl
import matplotlib.pyplot as plt
def plot_train():
print 'Plot data...'
data = read_csv('./train.csv', sep = ',')
for k in range(1, 5):
param = 'Wilderness_Area%s' % k
f = plt.figure(figsize = (8, 6))
p = data.pivot_table('Id', param, 'Cover_Type', 'count').plot(kind = 'barh', stacked = True, ax = f.gca())
img = './wilderness_area_cover_type_plot/Wilderness_Area%s_cover_type.png' % k
f.savefig(img)
for k in range(1, 41):
param = 'Soil_Type%s' % k
f = plt.figure(figsize = (8, 6))
p = data.pivot_table('Id', param, 'Cover_Type', 'count').plot(kind = 'barh', stacked = True, ax = f.gca())
img = './soil_type_cover_type_plot/Soil_Type%s_cover_type.png' % k
f.savefig(img)
def plot_elevation():
data = read_csv('./train.csv')
data = data.sort(['Elevation'])
print 'Plot Elevation...'
fig, axes = plt.subplots(ncols=1)
e = data.pivot_table('Id', ['Elevation'], 'Cover_Type', 'count').plot(ax=axes, title='Elevation')
f = e.get_figure()
f.savefig('./train_data_plot/elevation_cover_type.png')
def plot_aspect():
data = read_csv('./train.csv')
data = data.sort(['Aspect'])
print 'Plot Aspect...'
fig, axes = plt.subplots(ncols=1)
e = data.pivot_table('Id', ['Aspect'], 'Cover_Type', 'count').plot(ax=axes, title='Aspect')
f = e.get_figure()
f.savefig('./train_data_plot/aspect_cover_type.png')
def plot_slope():
data = read_csv('./train.csv')
data = data.sort(['Slope'])
print 'Plot Slope...'
fig, axes = plt.subplots(ncols=1)
e = data.pivot_table('Id', ['Slope'], 'Cover_Type', 'count').plot(ax=axes, title='Slope')
f = e.get_figure()
f.savefig('./train_data_plot/slope_cover_type.png')
def plot_horizontal_distance_to_hydrology():
data = read_csv('./train.csv')
data = data.sort(['Horizontal_Distance_To_Hydrology'])
print 'Plot Horizontal_Distance_To_Hydrology...'
fig, axes = plt.subplots(ncols=1)
e = data.pivot_table('Id', ['Horizontal_Distance_To_Hydrology'], 'Cover_Type', 'count').plot(ax=axes, title='Horizontal Distance To Hydrology')
f = e.get_figure()
f.savefig('./train_data_plot/horizontal_distance_to_hydrology_cover_type.png')
def plot_vertical_distance_to_hydrology():
data = read_csv('./train.csv')
data = data.sort(['Vertical_Distance_To_Hydrology'])
print 'Plot Vertical_Distance_To_Hydrology...'
fig, axes = plt.subplots(ncols=1)
e = data.pivot_table('Id', ['Vertical_Distance_To_Hydrology'], 'Cover_Type', 'count').plot(ax=axes, title='Vertical Distance To Hydrology')
f = e.get_figure()
f.savefig('./train_data_plot/vertical_distance_to_hydrology_cover_type.png')
def plot_horizontal_distance_to_roadways():
data = read_csv('./train.csv')
data = data.sort(['Horizontal_Distance_To_Roadways'])
print 'Plot Horizontal_Distance_To_Roadways...'
fig, axes = plt.subplots(ncols=1)
e = data.pivot_table('Id', ['Horizontal_Distance_To_Roadways'], 'Cover_Type', 'count').plot(ax=axes, title='Horizontal Distance To Roadways')
f = e.get_figure()
f.savefig('./train_data_plot/horizontal_distance_to_roadways_cover_type.png')
def plot_box():
data = read_csv("./train.csv")
headers = ['Elevation', 'Slope', 'Aspect', 'Horizontal_Distance_To_Hydrology', 'Vertical_Distance_To_Hydrology', 'Horizontal_Distance_To_Roadways']
headers += ['Hillshade_9am', 'Hillshade_Noon', 'Hillshade_3pm', 'Horizontal_Distance_To_Fire_Points']
for k in headers:
print "box plot %s..." % k.lower().replace("_", " ")
df = concat([data[k], data['Cover_Type']], axis=1, keys=[k, 'Cover_Type'])
f = plt.figure(figsize=(8, 6))
p = df.boxplot(by='Cover_Type', ax = f.gca())
f.savefig('./train_data_plot/box_%s_cover_type.png' % k.lower())
def get_train_data():
print 'Get train data...'
data = read_csv('./train.csv')
data = data.drop(['Id'], axis = 1)
# удаляем столбец Wilderness_Area2
data = data.drop(['Wilderness_Area2', 'Vertical_Distance_To_Hydrology', 'Slope'], axis = 1)
# data = data.drop(['Hillshade_9am', 'Hillshade_Noon', 'Hillshade_3pm'], axis = 1)
# удаляем столбцы SoilType1,...,SoilType40
drop_soil_type_cols = []
for k in range(1, 41):
cname = 'Soil_Type%s' % k
drop_soil_type_cols.append(cname)
data = data.drop(drop_soil_type_cols, axis = 1)
return data
def get_test_data():
print 'Get test data...'
data = read_csv('./test.csv')
result = DataFrame(data.Id)
# удаляем столбцы Id, Wilderness_Area2
data = data.drop(['Id', 'Wilderness_Area2', 'Vertical_Distance_To_Hydrology', 'Slope'], axis = 1)
# data = data.drop(['Hillshade_9am', 'Hillshade_Noon', 'Hillshade_3pm'], axis = 1)
# удаляем столбцы SoilType1,...,SoilType40
drop_soil_type_cols = []
for k in range(1, 41):
cname = 'Soil_Type%s' % k
drop_soil_type_cols.append(cname)
data = data.drop(drop_soil_type_cols, axis = 1)
return (data, result)
def cross_validation_test():
data = get_train_data()
target = data.Cover_Type
train = data.drop(['Cover_Type'], axis = 1)
kfold = 10
cross_val_final = {}
print 'Cross validation test...'
model_rfc = RandomForestClassifier(n_estimators = 1024, criterion='entropy', n_jobs = -1)
model_knc = KNeighborsClassifier(n_neighbors = 128)
model_lr = LogisticRegression(penalty='l1', C=1e5)
scores = cross_validation.cross_val_score(model_rfc, train, target, cv = kfold)
cross_val_final['RFC'] = scores.mean()
print 'RFC: ', scores.mean()
scores = cross_validation.cross_val_score(model_knc, train, target, cv = kfold)
cross_val_final['KNC'] = scores.mean()
print 'KNC: ', scores.mean()
scores = cross_validation.cross_val_score(model_lr, train, target, cv = kfold)
cross_val_final['LR'] = scores.mean()
print 'LR: ', scores.mean()
f = plt.figure(figsize = (8, 6))
p = DataFrame.from_dict(data = cross_val_final, orient='index').plot(kind='barh', legend=False, ax = f.gca())
f.savefig('./test_plot/cross_validation_rfc_1024.png')
# финальная функция
def go():
data = get_train_data()
model_rfc = RandomForestClassifier(n_estimators = 2500, criterion = 'entropy', n_jobs = -1)
# так как лучшие результаты у Random Forest
print 'Go!!!'
print 'RFC...'
test, result = get_test_data()
target = data.Cover_Type
test = test.drop(['Aspect'], axis = 1)
train = data.drop(['Cover_Type', 'Aspect'], axis = 1)
print "..."
model_rfc.fit(train, target)
result.insert(1,'Cover_Type', model_rfc.predict(test))
result.to_csv('./test_rfc_2500_new.csv', index=False)
def go_gbc():
data = get_train_data()
model_gbc = GradientBoostingClassifier(n_estimators = 1600)
print 'Go!!!'
print 'GBC...'
test, result = get_test_data()
target = data.Cover_Type
train = data.drop(['Cover_Type'], axis = 1)
model_gbc.fit(train, target)
result.insert(1,'Cover_Type', model_gbc.predict(test))
result.to_csv('./test_gbc_1600.csv', index=False)
def grid_search_test():
data = get_train_data()
target = data.Cover_Type
train = data.drop(['Cover_Type'], axis = 1)
model_rfc = RandomForestClassifier()
params = {"n_estimators" : [100, 250, 500, 625], "criterion" : ('entropy', 'gini')}
clf = grid_search.GridSearchCV(model_rfc, params)
clf.fit(train, target)
# summarize the results of the grid search
print(clf.best_score_)
print(clf.best_estimator_.criterion)
print(clf.best_estimator_.n_estimators)
# plot_elevation()
# plot_aspect()
# plot_slope()
# plot_horizontal_distance_to_hydrology()
# plot_vertical_distance_to_hydrology()
# plot_horizontal_distance_to_roadways()
# plot_train()
# cross_validation_test()
# grid_search_test()
# go()
# go_gbc()
# plot_box()
# data = get_train_data()
# train = data.drop(['Cover_Type', 'Aspect'], axis = 1)
# print train.head()