TODO: 训练精度更高的模型
This commit is contained in:
parent
68d7c5d2c9
commit
36b30982f7
40
data.py
40
data.py
|
|
@ -8,6 +8,7 @@ import pymysql
|
|||
import pickle
|
||||
import os
|
||||
import numpy
|
||||
import time, datetime
|
||||
|
||||
def get_collect():
|
||||
collect = {}
|
||||
|
|
@ -22,9 +23,10 @@ def get_collect():
|
|||
|
||||
# 使用 cursor() 方法创建一个游标对象 cursor
|
||||
cursor = db.cursor()
|
||||
|
||||
today = time.strftime("%Y-%m-%d", time.localtime())
|
||||
|
||||
# 使用 execute() 方法执行 SQL 查询
|
||||
print(cursor.execute('''SELECT coin, extra_coins, pay_users, create_at from pay_items_hour pih where region = "all" and platform="all"'''))
|
||||
cursor.execute('''SELECT coin, extra_coins, pay_users, create_at from pay_items_hour pih where region = "all" and platform="all" and create_at <= %s''', today)
|
||||
collect_pay = {}
|
||||
for row in cursor.fetchall():
|
||||
# print(row)
|
||||
|
|
@ -84,7 +86,7 @@ def get_collect():
|
|||
def load_pay_data(textNum = 80):
|
||||
|
||||
collect = get_collect()
|
||||
|
||||
|
||||
# TODO: 处理gift pay的波动关系
|
||||
|
||||
x_train = []
|
||||
|
|
@ -101,17 +103,20 @@ def load_pay_data(textNum = 80):
|
|||
total_coin = 0
|
||||
last_total_coin = 0
|
||||
|
||||
for v2 in lastday_v:
|
||||
last_total_coin += v2[0] + v2[1]
|
||||
|
||||
count = 0
|
||||
for v1, v2 in zip(cur_v,lastday_v):
|
||||
total_coin += v1[0] + v1[1]
|
||||
# print(v1[3])
|
||||
|
||||
last_total_coin += v2[0] + v2[1]
|
||||
# last_total_coin += v2[0] + v2[1]
|
||||
# print(v2[3])
|
||||
compare = float(total_coin - last_total_coin) / float(last_total_coin)
|
||||
# compare = float(total_coin - last_total_coin) / float(last_total_coin)
|
||||
# print(compare)
|
||||
|
||||
x_train.append([count, total_coin , (total_coin - last_total_coin) , v1[2] , v2[2]])
|
||||
x_train.append([count, total_coin, total_coin/last_total_coin])
|
||||
count+=1
|
||||
|
||||
for i in range(count):
|
||||
|
|
@ -119,15 +124,15 @@ def load_pay_data(textNum = 80):
|
|||
|
||||
lastday_v = cur_v
|
||||
|
||||
x_train = numpy.reshape(x_train, (len(x_train) , 5, 1))
|
||||
x_train = numpy.reshape(x_train, (len(x_train) , 3, 1))
|
||||
y_train = numpy.reshape(y_train, (len(y_train)))
|
||||
# max_features = 1024
|
||||
|
||||
tx_train = x_train[len(x_train) - textNum:]
|
||||
ty_train = y_train[len(y_train) - textNum:]
|
||||
|
||||
x_train = x_train[:len(x_train) - textNum]
|
||||
y_train = y_train[:len(y_train) - textNum]
|
||||
# x_train = x_train[:len(x_train) - textNum]
|
||||
# y_train = y_train[:len(y_train) - textNum]
|
||||
|
||||
return x_train, y_train, tx_train, ty_train
|
||||
|
||||
|
|
@ -150,23 +155,26 @@ def load_gift_data(textNum = 80):
|
|||
last_total_coin = 0
|
||||
users = 0
|
||||
|
||||
for v2 in lastday_v:
|
||||
last_total_coin += v2[0]
|
||||
|
||||
f = 20000000.0
|
||||
count = 0
|
||||
count = 1
|
||||
for v1, v2 in zip(cur_v,lastday_v):
|
||||
total_coin += v1[0]
|
||||
# print(v1[3])
|
||||
|
||||
last_total_coin += v2[0]
|
||||
# last_total_coin += v2[0]
|
||||
users += v1[1]
|
||||
|
||||
# print(v2[3])
|
||||
compare = float(total_coin - last_total_coin) / float(last_total_coin)
|
||||
# compare = float(total_coin - last_total_coin) / float(last_total_coin)
|
||||
# print(compare)
|
||||
|
||||
x_train.append([count, total_coin, compare, users ])
|
||||
x_train.append([count, total_coin, total_coin / last_total_coin, users ])
|
||||
count+=1
|
||||
|
||||
for i in range(count):
|
||||
for i in range(count - 1):
|
||||
y_train.append(total_coin)
|
||||
|
||||
lastday_v = cur_v
|
||||
|
|
@ -178,7 +186,7 @@ def load_gift_data(textNum = 80):
|
|||
tx_train = x_train[len(x_train) - textNum:]
|
||||
ty_train = y_train[len(y_train) - textNum:]
|
||||
|
||||
x_train = x_train[:len(x_train) - textNum]
|
||||
y_train = y_train[:len(y_train) - textNum]
|
||||
# x_train = x_train[:len(x_train) - textNum]
|
||||
# y_train = y_train[:len(y_train) - textNum]
|
||||
|
||||
return x_train, y_train, tx_train, ty_train
|
||||
24
predict.proto
Normal file
24
predict.proto
Normal file
|
|
@ -0,0 +1,24 @@
|
|||
syntax = "proto3";
|
||||
|
||||
service Predict {
|
||||
rpc PayDay(RequestPay) returns (Reply) {}
|
||||
rpc GiftDay(RequestGift) returns (Reply) {}
|
||||
}
|
||||
|
||||
// Request 统一Request请求
|
||||
message RequestPay {
|
||||
// count, total_coin , (total_coin - last_total_coin) , v1[2] , v2[2]] // 24小时
|
||||
int32 Hour = 1;
|
||||
int64 Coin = 2;
|
||||
int64 YesterdayCoin = 3;
|
||||
}
|
||||
|
||||
// RequestGift 统一Request请求
|
||||
message RequestGift {
|
||||
|
||||
}
|
||||
|
||||
// Reply 统一Reply 响应
|
||||
message Reply {
|
||||
|
||||
}
|
||||
|
|
@ -2,6 +2,9 @@ import numpy
|
|||
from keras.models import load_model
|
||||
from data import load_pay_data, load_gift_data
|
||||
|
||||
import matplotlib.pyplot as plt
|
||||
|
||||
|
||||
# x_train, y_train, tx_train, ty_train = load_pay_data(160)
|
||||
# model = load_model("./predict_pay")
|
||||
|
||||
|
|
@ -13,7 +16,6 @@ from data import load_pay_data, load_gift_data
|
|||
# print("测结果:", p_data[i][0], "测:", tx_train[i], "真实:", ty_train[i])
|
||||
|
||||
|
||||
|
||||
x_train, y_train, tx_train, ty_train = load_gift_data(160)
|
||||
model = load_model("./predict_gift")
|
||||
p_data = model.predict(tx_train)
|
||||
|
|
@ -23,6 +25,10 @@ for i in range(len(p_data)):
|
|||
if abs(comp) >= 0.1:
|
||||
print("测结果:", p_data[i][0], "测:", tx_train[i], "真实:", ty_train[i])
|
||||
|
||||
|
||||
plt.plot(ty_train)
|
||||
plt.plot(p_data)
|
||||
plt.show()
|
||||
# data = numpy.reshape([[15, 2359688 / 10000000, 255968 / 1000000, 10 / 10000]],(1, 4, 1))
|
||||
# print( model.predict(data))
|
||||
|
||||
|
|
|
|||
4
requirements.txt
Normal file
4
requirements.txt
Normal file
|
|
@ -0,0 +1,4 @@
|
|||
tensorflow
|
||||
keras
|
||||
numpy
|
||||
pymysql
|
||||
|
|
@ -18,23 +18,24 @@ if __name__ == "__main__":
|
|||
x_train, y_train, tx_train, ty_train = load_gift_data()
|
||||
|
||||
model = Sequential()
|
||||
units = 256
|
||||
units = 400
|
||||
|
||||
model.add(LSTM(units, activation='relu', input_shape=(4,1)))
|
||||
model.add(Dropout(0.3))
|
||||
model.add(LSTM(units, activation='relu', input_shape=(4,1) ))
|
||||
model.add(Dropout(0.2))
|
||||
|
||||
model.add(Dense(1))
|
||||
model.summary()
|
||||
|
||||
model.compile(loss='mse', optimizer='adam')
|
||||
|
||||
model.fit(x_train, y_train, batch_size=32, epochs=1000)
|
||||
model.fit(x_train, y_train, batch_size=1, epochs=50)
|
||||
model.save("./predict_gift")
|
||||
|
||||
p_data = model.predict(tx_train)
|
||||
for i in range(len(p_data)):
|
||||
comp = (p_data[i][0] - ty_train[i]) / ty_train[i]
|
||||
print(comp, p_data[i][0], ty_train[i])
|
||||
if abs(comp) >= 0.2:
|
||||
if abs(comp) >= 0.1:
|
||||
print("测结果:", p_data[i][0], "测:", tx_train[i], "真实:", ty_train[i])
|
||||
|
||||
|
||||
|
|
|
|||
10
train_pay.py
10
train_pay.py
|
|
@ -16,18 +16,18 @@ from data import load_pay_data
|
|||
|
||||
if __name__ == "__main__":
|
||||
|
||||
x_train, y_train, tx_train, ty_train = load_pay_data()
|
||||
x_train, y_train, tx_train, ty_train = load_pay_data(80)
|
||||
|
||||
|
||||
model = Sequential()
|
||||
units = 500
|
||||
model.add(LSTM(units, activation='relu', input_shape=(5,1)))
|
||||
model.add(Dropout(0.1))
|
||||
units = 500
|
||||
model.add(LSTM(units, activation='relu', input_shape=(3,1)))
|
||||
model.add(Dropout(0.3))
|
||||
model.add(Dense(1))
|
||||
model.summary()
|
||||
model.compile(loss='mse', optimizer='adam')
|
||||
|
||||
model.fit(x_train, y_train, batch_size=128, epochs=1500)
|
||||
model.fit(x_train, y_train, batch_size=1, epochs=50)
|
||||
model.save("./predict_pay")
|
||||
|
||||
p_data = model.predict(tx_train)
|
||||
|
|
|
|||
Loading…
Reference in New Issue
Block a user