/
anli.py
350 lines (319 loc) · 14.3 KB
/
anli.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
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
# -*- coding: utf-8 -*-
"""
Created on Thu Apr 19 11:29:50 2018
@author: Dell
"""
'''
ershoufang
'''
datalist=[]
with open(r'D:\workfile\work2\data\ershou\new.csv') as f:
for line in f:
line=line.strip(',,,,\n').split(',')
line=line[:5]+line[-1].split('/')
datalist.append(line)
import numpy as np
import pandas as pd
data=pd.DataFrame(datalist)
data=data.iloc[:,1:-1]
data.columns=['区','位置','描述','价格','小区','户型','面积','朝向','类型','电梯']
data.drop_duplicates()
#数据清洗
for indexs in data.index:
if(data.loc[indexs]['户型'] != None)and( '别墅' in data.loc[indexs]['户型']):
data.loc[indexs]['面积']=data.loc[indexs]['朝向']
data.loc[indexs]['朝向']=data.loc[indexs]['类型']
data.loc[indexs]['朝向']=data.loc[indexs]['类型']
data.loc[indexs]['类型']=data.loc[indexs]['电梯']
data.loc[indexs]['电梯']=None
for indexs in data.index:
if( data.loc[indexs].values[-1] == None) and (data.loc[indexs].values[2] != None)and ('电梯' in data.loc[indexs].values[2]):
data.loc[indexs].values[-1]='有电梯'
data=data.drop(data.index[data['户型']=='车位'])
for indexs in data.index:
if ( data.loc[indexs]['面积'] != None):
data.loc[indexs]['面积']=data.loc[indexs]['面积'][:-2]
for indexs in data.index:
if( data.loc[indexs]['价格'] != None):
data.loc[indexs]['价格']=data.loc[indexs]['价格'][:-1]
for indexs in data.index:
if(data.loc[indexs]['价格']!=None):
try:
data.loc[indexs]['价格']==int(data.loc[indexs]['价格'])
except ValueError:
data.loc[indexs]['价格']= None
data.info()
data=data.replace({None:-1})
data[['价格']]=data[['价格']].astype(int)
data[['面积']]=data[['面积']].astype(float)
data=data.replace({-1:np.NaN})
#出售量
df_district = data['区'].value_counts()
failegetdata=df_district[df_district <2]
for indexs in data.index:
if data.loc[indexs]['区'] in list(failegetdata.index):
data=data.drop([indexs])
df_district = data['区'].value_counts()
import matplotlib.pyplot as plt
plt.rcParams['font.sans-serif']=['SimHei']
plt.rcParams['axes.unicode_minus'] = False
df_district.plot(kind="bar", title="各区在售二手房数量",figsize = (12, 9))
df_xq = data['小区'].value_counts()
df_xq[:5].plot(kind="bar", title="在售数量最多的五个小区",figsize = (12, 9))
#出售价
data.to_csv('tmp.csv')
data=pd.read_csv('tmp.csv',encoding='gbk')
data['单价']=data['价格']/data['面积']
data['单价'][data['单价']>0.3].plot.hist(bins = 20,title='在售二手房单价直方图',figsize = (12, 9))
a=data[['单价','区']].dropna()
grouped =a['单价'].groupby(a['区']).apply(np.median)
a=list(grouped.index[0:2]+'区')
a.append('亦庄开发区')
a=a+list(grouped.index[3:]+'区')
grouped.index=a
from pyecharts import Map
value =list(grouped)
attr =list(grouped.index)
map=Map("二手房在售单价中位数 单位:万元", width=1200, height=600)
map.add("", attr, value, maptype='北京', is_visualmap=True, visual_text_color='#404a59',visual_range=[1,10],is_map_symbol_show =False)
map.show_config()
map.render()
df=pd.read_csv('ershoufangshuju.csv',encoding='gbk')
del df['Unnamed: 0']
df['电梯'][df['类型']=='有电梯']=df['类型'][df['类型']=='有电梯']
df['电梯'][df['类型']=='无电梯']=df['类型'][df['类型']=='无电梯']
df['类型'][df['类型']=='有电梯']=np.NaN
df['类型'][df['类型']=='无电梯']=np.NaN
tmp=pd.crosstab(df['区'],df['类型'],margins=True)
def plotbar_Stacked(tmp,t=''):
tmp=tmp.apply(lambda x:x/x[-1],axis=1)
tmp=tmp.drop('All',1).drop('All',0)
tmp.plot.bar(stacked=True,figsize = (12, 9),title=str(t))
plotbar_Stacked(tmp,'各区在售二手房装修类型比重')
df['室数']=df['户型']
#for indexs in df.index:
#if(df.loc[indexs]['户型'] is not np.NaN):
#df.loc[indexs]['室数']=df.loc[indexs]['户型'][0]
df['室数']=df['室数'].apply(lambda x: x[0] if not x is np.NaN else x)
df['室数']=df['室数'].apply(lambda x: '别墅' if x in ['双','叠','独','联'] else x)
df['室数']=df['室数'].apply(lambda x: '3居室以下' if x in ['1','2','3'] else x)
df['室数']=df['室数'].apply(lambda x: '4-6居室' if x in ['4','5','6'] else x)
df['室数']=df['室数'].apply(lambda x: '别墅' if x in ['7','8','9'] else x)
tmp=pd.crosstab(df['区'],df['室数'],margins=True)
df=data.drop(df.index[df['室数']=='车'])
plotbar_Stacked(tmp,'各区在售二手房室数比重')
import seaborn as sns
a=df[['价格','类型','区']].dropna()
a=a.pivot_table(index='区',columns='类型',values='价格', aggfunc=np.median)
sns.factorplot(x='类型', y='价格',data=df, size=4, aspect=2)
a=a.ix[[0,1,9,10,11,12],]
a.plot.line(stacked=False,color=['yellow','red','blue','green'],figsize = (12, 9),title='城六区各装修类型二手房售价中位数',fro)
df1=pd.read_csv(r'D:\workfile\work1\bj_info.csv')
del df1['Unnamed: 0']
df1=df1[df1['参考均价']!='暂无报价']
import json
from urllib.request import urlopen, quote
import requests,csv
import pandas as pd #导入这些库后边都要用到
def getlnglat(address):
url = 'http://api.map.baidu.com/geocoder/v2/'
output = 'json'
ak = 'ry6kSSLorwwNZGUKPe3hZUjSDmyI1maC'
add = quote(address) #由于本文城市变量为中文,为防止乱码,先用quote进行编码
uri = url + '?' + 'address=' + add + '&output=' + output + '&ak=' + ak
req = urlopen(uri)
res = req.read().decode() #将其他编码的字符串解码成unicode
temp = json.loads(res) #对json数据进行解析
return temp
file = open('point_f.json','w')
faillist=[]
for indexs in df1.index:
a=df1.loc[indexs]['地址']
b=df1.loc[indexs]['参考均价']
try:
lng = getlnglat(a)['result']['location']['lng']
lat = getlnglat(a)['result']['location']['lat']
str_temp = '{"lat":' + str(lat) + ',"lng":' + str(lng) + ',"count":' + str(b) +'},'
except KeyError as e:
faillist.append(a)
print(e)
file.write(str_temp)
file.close()
#建模
f = open(r'D:\workfile\work1\二手房在售数据整理(纯数字)\二手房在售数据整理(纯数字).csv')
df=pd.read_csv(f)
#数据预处理
from sklearn import preprocessing
df['面积_scale']=preprocessing.scale(df['面积'])
df['卧室'][df['卧室']<=3]=0
df['卧室'][df['卧室']>3]=1
df['客厅'][df['客厅']<2]=0
df['客厅'][df['客厅']>=2]=1
y=df.ix[:,0]
X=df.ix[:,2:]
from sklearn.ensemble import AdaBoostRegressor
#特征选择
# AdaBoost
ada_est =AdaBoostRegressor(random_state=0)
ada_param_grid = {'n_estimators': [500], 'learning_rate': [0.01, 0.1]}
ada_grid = model_selection.GridSearchCV(ada_est, ada_param_grid, n_jobs=25, cv=10, verbose=1)
ada_grid.fit(X, y)
print('Top N Features Best Ada Params:' + str(ada_grid.best_params_))
print('Top N Features Best Ada Score:' + str(ada_grid.best_score_))
print('Top N Features Ada Train Score:' + str(ada_grid.score(X, y)))
feature_imp_sorted_ada = pd.DataFrame({'feature': list(X),
'importance': ada_grid.best_estimator_.feature_importances_}).sort_values('importance', ascending=False)
features_top_n_ada = feature_imp_sorted_ada.head(7)['feature']
print(str(features_top_n_ada))
del df['朝向']
tmp=pd.get_dummies(df['装修'])
df=pd.concat([df, tmp], axis=1)
del df['装修']
y=df.ix[:,0]
X=df.ix[:,2:]
#朝向重要性 0,舍去
from sklearn import ensemble
from sklearn import model_selection
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)
#调参
#gbm
gbm_reg = GradientBoostingRegressor(random_state=42)
gbm_reg_param_grid = {'n_estimators': [2000], 'max_depth': [4], 'learning_rate': [0.01,0.2], 'max_features': [3]}
gbm_reg_grid = model_selection.GridSearchCV(gbm_reg, gbm_reg_param_grid, cv=10, n_jobs=25, verbose=1, scoring='neg_mean_squared_error')
gbm_reg_grid.fit(X_train, y_train)
print(' Best Params:' + str(gbm_reg_grid.best_params_))
gbm_reg_grid=GradientBoostingRegressor(n_estimators=2000,max_depth=4, learning_rate= 0.01, max_features=3)
gbm_reg_grid.fit(X_train, y_train)
from sklearn.externals import joblib
y_predict_gbm=gbm_reg_grid.predict(X_test)
mae_gbm=(np.abs(y_predict_gbm-y_test)).sum()/9467
joblib.dump(gbm_reg_grid, 'gbm_reg_grid.model')
print(mae_gbm)
#mae=1.1141107597511322
#rf
rf_reg = RandomForestRegressor()
rf_reg_param_grid = {'n_estimators': [2000], 'max_depth': [5,10], 'random_state': [0]}
rf_reg_grid = model_selection.GridSearchCV(rf_reg, rf_reg_param_grid, cv=10, n_jobs=25, verbose=1, scoring='neg_mean_squared_error')
rf_reg_grid.fit(X_train, y_train)
print(' Best Params:' + str(rf_reg_grid.best_params_))
rf_reg_grid=RandomForestRegressor(n_estimators=2000,max_depth=10, random_state=0)
rf_reg_grid.fit(X_train, y_train)
joblib.dump(rf_reg_grid, 'rf_reg_grid.model')
y_predict_rf=rf_reg_grid.predict(X_test)
mae_rf=(np.abs(y_predict_rf-y_test)).sum()/9467
print(mae_rf)
#mae=1.1062916504558062
#ada (舍去)
ada_est_param_grid = {'n_estimators': [5000], 'learning_rate': [0.02,0.1], 'random_state': [0]}
ada_est =AdaBoostRegressor()
ada_est_grid = model_selection.GridSearchCV(ada_est, ada_est_param_grid, cv=10, n_jobs=25, verbose=1, scoring='neg_mean_squared_error')
ada_est_grid.fit(X_train, y_train)
print(' Best Params:' + str(ada_est_grid.best_params_))
joblib.dump(rf_reg_grid, 'ada_est.model')
ada_est=AdaBoostRegressor(n_estimators=5000,learning_rate=0.02)
ada_est.fit(X_train, y_train)
y_predict_ada=ada_est.predict(X_test)
mae_ada=(np.abs(y_predict_ada-y_test)).sum()/9467
print(mae_ada)
#1.2866268400664551
#svr
from sklearn.svm import SVR
svr_param_grid = {'kernel':('rbf','linear')}
svr = SVR(gamma=0.1)
svr_grid = model_selection.GridSearchCV(svr, svr_param_grid, cv=10, n_jobs=25, verbose=1, scoring='neg_mean_squared_error')
svr_grid.fit(X_train, y_train)
print(' Best Params:' + str(svr_grid.best_params_))
svr = SVR(gamma=0.1,kernel='rbf')
svr.fit(X_train, y_train)
y_predict_svr=svr.predict(X_test)
mae_svr=(np.abs(y_predict_svr-y_test)).sum()/9467
joblib.dump(svr, 'svr.model')
print(mae_svr)
#KNN
from sklearn.neighbors import KNeighborsRegressor
KNN = KNeighborsRegressor()
knn_param_grid = {'n_neighbors':[3,10]}
knn_grid = model_selection.GridSearchCV(KNN, knn_param_grid, cv=10, n_jobs=25, verbose=1, scoring='neg_mean_squared_error')
knn_grid.fit(X_train, y_train)
print(' Best Params:' + str(knn_grid.best_params_))
KNN = KNeighborsRegressor(n_neighbors=10)
KNN.fit(X_train, y_train)
y_predict_knn=KNN.predict(X_test)
mae_knn=(np.abs(y_predict_knn-y_test)).sum()/9467
joblib.dump(KNN, 'KNN.model')
print(mae_knn)
#mlp
from sklearn.neural_network import MLPRegressor
MLP = MLPRegressor(hidden_layer_sizes=(300, 200,200),max_iter=100,activation='relu')
MLP.fit(X_train, y_train)
y_predict_MLP=MLP.predict(X_test)
mae_MLP=(np.abs(y_predict_MLP-y_test)).sum()/9467
joblib.dump(MLP, 'MLP.model')
print(mae_MLP)
#xgb
import xgboost as xgb
x_regress = xgb.XGBRegressor(max_depth=20,n_estimators =5000)
x_regress_param_grid = {'max_depth': [5,20]}
x_regress_grid = model_selection.GridSearchCV(x_regress, x_regress_param_grid, cv=10, n_jobs=25, verbose=1, scoring='neg_mean_squared_error')
x_regress.fit(X_train, y_train)
joblib.dump(x_regress, 'x_regress_grid.model')
y_predict_xgb=x_regress.predict(X_test)
mae_xgb=(np.abs(y_predict_xgb-y_test)).sum()/9467
# 模型融合
#简单平均
pred=pd.DataFrame({'ada':y_predict_ada,'gbm':y_predict_gbm,'rf':y_predict_rf,'svr':y_predict_svr,'knn':y_predict_knn,'mlp':y_predict_MLP,'xgb':y_predict_xgb})
pred['pred']=pred['ada']+pred[ 'knn']+pred[ 'gbm']+pred['mlp']+pred['rf']+pred['svr']+pred['xgb']
pred['pred']=pred['pred']/7
mae_final=(np.abs(pred['pred']-y_test)).sum()/9467
print(mae_final)
#0.8367033561074727
pred_['pred']=pred[ 'gbm']+pred['mlp']+pred['rf']+pred['svr']+pred['xgb']+pred['knn']
pred['pred']=pred['pred']/6
mae_final_1=(np.abs(pred['pred']-y_test)).sum()/9467
print(mae_final_1)
#加权平均
pred['pred']=0.332020866*pred[ 'gbm']+0.334367542*pred['rf']+0.333611592*pred['svr']
pred['pred']=pred['pred']/3
mae_final_1=(np.abs(pred['pred']-y_test)).sum()/9467
print(mae_final_1)
#stacking
from sklearn.model_selection import KFold
ntrain = X_train.shape[0]
ntest = X_test.shape[0]
SEED = 0 # for reproducibility
NFOLDS = 3 # set folds for out-of-fold prediction
kf = KFold(n_splits = NFOLDS, random_state=SEED, shuffle=False)
def get_out_fold(clf, x_train, y_train, x_test):
oof_train = np.zeros((ntrain,))
oof_test = np.zeros((ntest,))
oof_test_skf = np.empty((NFOLDS, ntest))
for i, (train_index, test_index) in enumerate(kf.split(x_train)):
x_tr = x_train[train_index]
y_tr = y_train[train_index]
x_te = x_train[test_index]
clf.fit(x_tr, y_tr)
oof_train[test_index] = clf.predict(x_te)
oof_test_skf[i, :] = clf.predict(x_test)
oof_test[:] = oof_test_skf.mean(axis=0)
return oof_train.reshape(-1, 1), oof_test.reshape(-1, 1)
x_train = X_train.values
x_test = X_test.values
y_train_1 = y_train.values
gbm_oof_train, gbm_oof_test = get_out_fold(gbm_reg_grid, x_train, y_train_1, x_test)
rf_oof_train, rf_oof_test = get_out_fold(rf_reg, x_train, y_train_1, x_test)
#ada_oof_train, ada_oof_test = get_out_fold(ada_est, x_train, y_train_1, x_test)
svr_oof_train,svr_oof_test = get_out_fold(svr, x_train, y_train_1, x_test)
#knn_oof_train, bayes_oof_test = get_out_fold(KNN, x_train, y_train_1, x_test)
#MLP_oof_train, MLP_oof_test = get_out_fold(MLP, x_train, y_train_1, x_test)
#xbg_oof_train, xbg_oof_test = get_out_fold(x_regress, x_train, y_train_1, x_test)
knn_oof_test=bayes_oof_test
#gbt LEVEL2
x_train_f = np.concatenate((gbm_oof_train, rf_oof_train, svr_oof_train), axis=1)
x_test_f = np.concatenate((gbm_oof_test, rf_oof_test, svr_oof_test), axis=1)
gbm_f = GradientBoostingRegressor(random_state=42,n_estimators=2000,max_depth=5,learning_rate=0.01,max_features=3)
gbm_f.fit(x_train_f, y_train)
y_predict_f=gbm_f.predict(x_test_f)
mae_gbm_f=(np.abs(y_predict_f-y_test)).sum()/9467