forked from 18786683795/LSTM_realTmpredict_rhyc
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathTimeSeries_predict_rh.py
250 lines (215 loc) · 9.87 KB
/
TimeSeries_predict_rh.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
# -*- coding: utf-8 -*-
"""
Created on Tue Aug 20 09:54:18 2019
@author: Lenovo
"""
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
import os
import sklearn.metrics as metrics
import datetime
import time
import schedule
import pymysql
from sqlalchemy import create_engine
from data_processing import data_processing
from datetime import datetime, timedelta
#定义超参数
input_size=13
output_size=13
mean=np.array([358.637,13.5976,80.174,16.5052,25.3909,60.5588,9.5699,11.0471,36.8855,6.2935,27.0285,50.9569,20.6291])
std=np.array([362.02239963,26.59734713,92.56434046,25.6864239,51.21534631,71.97984122,21.45158069,18.20996105,51.09030231,17.70573799,36.89085642,74.19714713,34.06744095])
#def load_data():
# print("开始准备特征数据：》》》》》》》》》》》")
# df=data_processing()
# print("数据准备完毕：！！！")
# time.sleep(5)
# print("开始交通流量预测：》》》》》》》》》》》》》》》》》")
# return df
print("开始准备特征数据：》》》》》》》》》》》")
df=data_processing()
print("数据准备完毕：！！！")
time.sleep(2)
print("开始交通流量预测：》》》》》》》》》》》》》》》》》")
def Lstm_forecast(rnn_unit,lr,md):
tf.reset_default_graph()
def forecast_Model(lk):
# df=load_data()
col = df.columns
df1=df[df['lk']==lk]
data=df1[col[3:16]].astype('int').values
print(lk+"数据准备完毕：！！！")
tf.reset_default_graph()
def get_forecast_data(time_step=6,test_begin=0):
data_test=data[test_begin:]
# mean=np.mean(data_test,axis=0)
# std=np.std(data_test,axis=0)
normalized_test_data=(data_test-mean)/std #标准化
size=(len(normalized_test_data)+time_step-1)//time_step
test_x,test_y=[],[]
for i in range(size-1):
x=normalized_test_data[i:i+time_step,:13]
y=normalized_test_data[i:i+time_step,:13]
test_x.append(x.tolist())
test_y.extend(y)
# test_x.append((normalized_test_data[(i+1)*time_step:,:5]).tolist())
# test_y.extend((normalized_test_data[(i+1)*time_step:,:5]).tolist())
return mean,std,test_x,test_y
#定义权重及偏差
weights={
'in':tf.Variable(tf.random_normal([input_size,rnn_unit])),
'out':tf.Variable(tf.random_normal([rnn_unit,13]))
}
biases={
'in':tf.Variable(tf.constant(0.1,shape=[rnn_unit,])),
'out':tf.Variable(tf.constant(0.1,shape=[13,]))
}
def lstm(X):
batch_size=tf.shape(X)[0]
time_step=tf.shape(X)[1]
w_in=weights['in']
b_in=biases['in']
input=tf.reshape(X,[-1,input_size])
input_rnn=tf.matmul(input,w_in)+b_in
input_rnn=tf.reshape(input_rnn,[-1,time_step,rnn_unit])
cell=tf.nn.rnn_cell.BasicLSTMCell(rnn_unit,reuse=tf.AUTO_REUSE)
init_state=cell.zero_state(batch_size,dtype=tf.float32)
output_rnn,final_states=tf.nn.dynamic_rnn(cell, input_rnn,initial_state=init_state, dtype=tf.float32)
output=tf.reshape(output_rnn,[-1,rnn_unit])
w_out=weights['out']
b_out=biases['out']
pred=tf.matmul(output,w_out)+b_out
return pred,final_states
def prediction(time_step=6):
print("开始交通流量预测：》》》》》》》》》》》")
X=tf.placeholder(tf.float32, shape=[None,time_step,input_size])
#Y=tf.placeholder(tf.float32, shape=[None,time_step,output_size])
_,_,test_x,test_y=get_forecast_data(time_step)
pred,_=lstm(X)
saver=tf.train.Saver(tf.global_variables())
with tf.Session() as sess:
#参数恢复
base_path = os.getcwd()
module_file = tf.train.latest_checkpoint(base_path+md)
saver.restore(sess, module_file)
#取训练集最后一行为测试样本。shape=[1,time_step,input_size]
prev_seq=test_x[-1]
# print("prev_seq1:",prev_seq)
predict=[]
#得到之后100个预测结果
for i in range(1):
next_seq=sess.run(pred,feed_dict={X:[prev_seq]})
predict.append(pd.DataFrame(next_seq[-1]).T)
#每次得到最后一个时间步的预测结果，与之前的数据加在一起，形成新的测试样本
prev_seq=np.vstack((prev_seq[1:],next_seq[-1]+((np.random.uniform(-5,8,[1,13])-mean)/std)))
# rel_predict=predict[-1]*std[:5]+mean[:5]
rel_predict = []
for i in predict*std[:13]+mean[:13]:
predict1 = i.reshape(13)
rel_predict.append(abs(predict1))
rel_predict=round(pd.DataFrame(rel_predict))
dat = data.mean(axis=0)
f = pd.DataFrame((abs(rel_predict).values-dat)/dat).replace(np.inf,0)
f = f.replace(np.nan,0).convert_objects(convert_numeric=True)
rel_predict2 = pd.DataFrame(np.where(f>0.05,dat,rel_predict))
g = pd.DataFrame((abs(rel_predict2).values-dat)/dat).replace(np.inf,0)
g = g.replace(np.nan,0).convert_objects(convert_numeric=True)
rel_predict1 = pd.DataFrame(np.where(g<-0.05,dat,rel_predict2))
# fac = (abs(rel_predict).values-dat)/dat
# func = [lambda fac: 0.15 < fac < 3.5 and -3.5 < fac < -0.15]
# rel_predict1 = round(pd.DataFrame(np.where(func,dat,rel_predict)))
# f=rel_predict1.values
# func1=[lambda f: f==0]
# rel_predict2 = round(pd.DataFrame(np.where(func1,rel_predict1,rel_predict)))
return rel_predict1
rel_predict = prediction()
return rel_predict
def main():
today = datetime.now()-timedelta(days=1)
#Tm = (today + timedelta(minutes=10)).strftime("%Y-%m-%d %H:%M")
Tm = today.strftime("%Y-%m-%d %H:%M")
sd = Tm[14:16]
if int(sd) in range(10):
t=10
if int(sd) in range(10,20):
t=20
if int(sd) in range(20,30):
t=30
if int(sd) in range(30,40):
t=40
if int(sd) in range(40,50):
t=50
if int(sd) in range(50,60):
t=60
prediction = []
for lk in set(df['lk']):
yc = forecast_Model(lk)
col=df.columns
yc.columns = col[3:16]
yc['lk'] = lk
yc['sj'] = Tm[:14]+str(t)
yc['minut'] = range(1) ##############
cols = list(yc)
cols.insert(0,cols.pop(cols.index('lk')))
cols.insert(1,cols.pop(cols.index('sj')))
cols.insert(2,cols.pop(cols.index('minut')))
yc = yc.loc[:,cols]
prediction.append(yc)
prediction = pd.concat(prediction)
# print(prediction)
predt = []
for lk in set(prediction['lk']):
dt1 = prediction[prediction['lk']==lk]
sj2 = [x[:14] for x in dt1['sj']]
# sj3 = np.array([str(t+x) for x in range(6)])
# sj1 = sj2+(dt1['minut']+1).astype('str')+str(0)
# dt1['sj'] = sj1 ###6个时刻
dt1['sj'] = yc['sj'] ####1个时刻
predt.append(dt1)
predt = pd.concat(predt)
# print(predt)
print("交通流量预测完毕：》》》》》》》》》》》")
return predt
predt0=main()
return predt0
#预测结果写入数据库
#engine=create_engine("mysql+pymysql://root:[email protected]:3306/keenIts?charset=utf8")
#predt.to_sql(name='flow_TimeSeries',con=engine,if_exists='append',index=False,index_label=False)
def forecast_main():
predict1 = Lstm_forecast(rnn_unit=10,lr=0.0006,md='/model_10_0.0006/')
tf.reset_default_graph()
predict2 = Lstm_forecast(rnn_unit=15,lr=0.0008,md='/model_15_0.0008/')
tf.reset_default_graph()
predict3 = Lstm_forecast(rnn_unit=30,lr=0.0005,md='/model_30_0.0005/')
result = round(predict1[predict1.columns[3:]]*0.35+predict2[predict2.columns[3:]]*0.35+predict3[predict3.columns[3:]]*0.3)
result.columns = predict1.columns[3:]
result[['lk','sj','minut']] = predict1[['lk','sj','minut']]
#交换列顺序
cols = list(result)
cols.insert(0,cols.pop(cols.index('lk')))
cols.insert(1,cols.pop(cols.index('sj')))
cols.insert(2,cols.pop(cols.index('minut')))
result = result.loc[:,cols]
# print("RESULT:",result)
result.to_csv('result.csv')
result1=result.drop(['minut'],axis=1)
result1['point_number']=result['lk']
result1['time_part']=[x[11:] for x in result1['sj']]
result1['date']=[x[:10] for x in result1['sj']]
result2=result1.drop(['sj','lk'],axis=1)
#交换列顺序
cols = list(result2)
cols.insert(0,cols.pop(cols.index('point_number')))
cols.insert(1,cols.pop(cols.index('date')))
cols.insert(2,cols.pop(cols.index('time_part')))
result2 = result2.loc[:,cols]
result2 = result2.sort_values(by=['point_number'])
print("RESULT:",result2)
print("预测结果写入数据库》》》》》》》》》》》")
#预测结果写入数据库
engine=create_engine("mysql+pymysql://root:[email protected]:3306/keenIts?charset=utf8")
result2.to_sql(name='flow_forecast_TimeSeries',con=engine,if_exists='append',index=False,index_label=False)
print("预测完毕！！！！！！！！！！！")
forecast_main()