股市预测python代码<循环神经网络>

数据集：（百度网盘）
链接：https://pan.baidu.com/s/1ozBoPhO7LU06z9mNlufbmg
提取码：59vk

#基于。csv数据没建立rnn模型，预测
# 1.数据预处理，将序列数据转化为可用于rnn输入的数据
# 2.对新数据进行预测（预测数据集），可视化结果
# 3.存储预测结果（csv文件），并观察局部预测结果（局限性）
#模型结构：单纯rnn，输出有五个神经元，每次使用前8个数据预测第9个数据

import pandas as pd
import numpy as np
data = pd.read_csv('D:\\pydata\\zgpa_train.csv')

data.head()
price = data.loc[:,'close']
price.head()

#归一化处理
price_norm = price/max(price)
print(price_norm)

#可视化
from matplotlib import pyplot as plt
fig1=plt.figure(figsize=(8,5))#建立图形
plt.plot(price)#画图
#加标注
plt.title('close price')
plt.xlabel('time')
plt.ylabel('price')
plt.show()

#xy的赋值

#define X and y
#define method to extract X and y

#提取数据序列
def extract_data(data,time_step):
    X=[]
    y=[]
    #0,1,2,3...9:10个样本；time_step=8;0,1,2..7;12,3,...9
    for i in range(len(data)-time_step):
        X.append([a for a in data[i:i+time_step]])#X数组是训练集
        y.append(data[i+time_step])#y是预测集
    X = np.array(X)#转换为一个数组
    X = X.reshape(X.shape[0],X.shape[1],1)
    return X,y

time_step = 8
X,y = extract_data(price_norm,time_step)
print("Xshape:")
print(X.shape)
print(y)

from keras.models import Sequential
from keras.layers import Dense,SimpleRNN
model = Sequential()

#单层有五个神经元units，input_shape(样本数，序列长度，序列维度）:数据是一维的，activation：激活函数
model.add(SimpleRNN(units = 5, input_shape=(time_step,1),activation='relu'))
#输出数值：索引神经元为1；线性（回归预测常用）
#输出y的一层
model.add(Dense(units=1,activation='linear'))

#configure the model
#(优化器，损失函数）
#平方差mean_squared_errar
model.compile(optimizer='adam',loss='mean_squared_error')
model.summary()

#模型训练
#打印X，y的维度
#print(X.shape(X),len(y))
#训练样本，每次30个样本，共训练200次
y=np.array(y);
model.fit(X,y,batch_size=30,epochs=200)


#预测结果可视化
y_train_predict = model.predict(X)*max(price)
#y_train = [i*max(price) for i in y]#把归一化之后的数值转换过来
y_train = y*max(price)
print(y_train_predict,y_train)

fig2=plt.figure(figsize=(8,5))#建立图形
plt.plot(y_train,label='real price')#画图
plt.plot(y_train_predict,label='predict price')#画图
#加标注
plt.title('close price')
plt.xlabel('time')
plt.ylabel('price')
plt.legend()
plt.show()


#predict

data_test = pd.read_csv('D:\\pydata\\zgpa_test.csv')
data_test.head()
price_test = data_test.loc[:,'close']
price_test.head()
price_test_norm = price_test/max(price)#归一化
X_test_norm,y_test_norm = extract_data(price_test_norm,time_step)
print(X_test_norm.shape,len(y_test_norm))

y_test_predict = model.predict(X_test_norm)*max(price)
y_test_norm=np.array(y_test_norm);
y_test = y_test_norm*max(price)

fig3=plt.figure(figsize=(8,5))#建立图形
plt.plot(y_test,label='real price test')#画图
plt.plot(y_test_predict,label='predict price test')#画图
#加标注
plt.title('close price')
plt.xlabel('time')
plt.ylabel('price')
plt.legend()
plt.show()

for i in range(0,len(y_test_predict)):
    totforenum=abs(y_test_predict[i]-y_test[i]);
    totnum=y_test[i];
totnum/=len(y_test_predict);
totforenum/=len(y_test_predict);
print("平均误差率:",end="");
print(totforenum/totnum*100,end="");
print("%");


#存储数据
result_y_test = np.array(y_test).reshape(-1,1)
result_y_test_predict = y_test_predict
print(result_y_test.shape,result_y_test_predict.shape)
result = np.concatenate((result_y_test,result_y_test_predict),axis = 1)
print(result.shape)
result = pd.DataFrame(result,columns=['real_price_test','predict_price_test'])
result.to_csv('zapa_predict_test.csv')

多跑几次的预测结果不一样，神经网络好奇妙嗷

 这是两年前写的一个代码了，当时还是比较稚嫩，代码水平比较青涩，很多学大数据的同学可能会用得到，这里强烈建议学习python的同学使用jupyter，它的每个代码段之间的运行结果是保存的，相当于断点，非常方便调试。

运行结果：

 代码：

import tushare  as ts
import numpy as np
import pandas as pd
from sklearn.linear_model import LinearRegression
import matplotlib.pyplot as plt
import matplotlib as mpl
pro=ts.pro_api('cf405b27c3824376311bb83ebd0f5217cc504283f4d235ae27427521')
df = pro.daily(ts_code='600020.SH', start_date='20190908', end_date='20191120')
date=df.trade_date
close=df.close
close=close.values.reshape(-1,1)
Short​import tushare  as ts
import numpy as np
import pandas as pd
from sklearn.linear_model import LinearRegression
import matplotlib.pyplot as plt
import matplotlib as mpl
pro=ts.pro_api('cf405b27c3824376311bb83ebd0f5217cc504283f4d235ae27427521')
df = pro.daily(ts_code='600020.SH', start_date='20190908', end_date='20191120')
date=df.trade_date
close=df.close
close=close.values.reshape(-1,1)
Short=12
Long=26
Mid=9
b=len(close)
a=close.copy()
date1=[]
close1=[]
for i in range (b):
    close[i]=a[b-i-1]
    if i >25:
        d=a[b-i-1]
        close1.append(d)
        c=date[b-1-i]
        date1.append(c)
def EMA(a,N):
    ema=a.copy()
    b=len(a)
    for i in range(b):
        if i == 0 :
            ema[i]=a[i];
        if i > 0 :
            ema[i]=(ema[i-1]*(N-1)+a[i]*2)/(N+1)
    return ema
DIF=EMA(close,Short)-EMA(close,Long)
DEM=EMA(DIF,Mid)
MACD=(DIF-DEM)*2
MACD1=[]
for i in range(b):
    if i >25:
        e=MACD[i]
        MACD1.append(e)
mpl.rcParams['font.sans-serif'] = ['KaiTi']
mpl.rcParams['font.serif'] = ['KaiTi']
mpl.rcParams['axes.unicode_minus'] = False 
model=LinearRegression()
model.fit(close1,MACD1)
plt.figure(figsize=(15,5))
plt.title('MACD of the 600020.SH\'s stock')
plt.plot(date1,MACD1,'r-',label='实际')
plt.plot(date1,model.predict(close1),'b--',label='预测值')
plt.legend()=12
Long=26
Mid=9
b=len(close)
a=close.copy()
date1=[]
close1=[]
for i in range (b):
    close[i]=a[b-i-1]
    if i >25:
        d=a[b-i-1]
        close1.append(d)
        c=date[b-1-i]
        date1.append(c)
def EMA(a,N):
    ema=a.copy()
    b=len(a)
    for i in range(b):
        if i == 0 :
            ema[i]=a[i];
        if i > 0 :
            ema[i]=(ema[i-1]*(N-1)+a[i]*2)/(N+1)
    return ema
DIF=EMA(close,Short)-EMA(close,Long)
DEM=EMA(DIF,Mid)
MACD=(DIF-DEM)*2
MACD1=[]
for i in range(b):
    if i >25:
        e=MACD[i]
        MACD1.append(e)
mpl.rcParams['font.sans-serif'] = ['KaiTi']
mpl.rcParams['font.serif'] = ['KaiTi']
mpl.rcParams['axes.unicode_minus'] = False 
model=LinearRegression()
model.fit(close1,MACD1)
plt.figure(figsize=(15,5))
plt.title('MACD of the 600020.SH\'s stock')
plt.plot(date1,MACD1,'r-',label='实际')
plt.plot(date1,model.predict(close1),'b--',label='预测值')
plt.legend()