order ={"0":"Open","1":"High","2":"Low","3":"Close","4":"Volume*e-6"}
yahooData = np.load('yahooData.npy')
Adj_Close,High,Low,Close,Open = np.hsplit(yahooData,5)
Volume = Adj_Close
yahooData = np.hstack([Open,High,Low,Close,Volume])
shape = yahooData.shape
print "shape = ",shape
print "-"*80
# print yahooData[0:10,:]
print "Close shape = ",Close.shape
myNNmodel = MyNeurolNetworkModel()
kdj = myNNmodel.calculate_kdj(yahooData) #(None,3)
print "kdj = ",kdj.shape
declose = myNNmodel.calculate_dclose(Close) #(None,4)
print "declose = ",declose.shape
logfit = myNNmodel.calculate_logfit(Close) #(None,1)
print "logfit = ",logfit.shape
closeExponent = myNNmodel.calculate_exponent(Close,exponent = 0.9) #(None,1)
print "closeExponent = ",closeExponent.shape
from sklearn.decomposition import PCA
pca = PCA(n_components = 2)
# x_train = x_sample[train_start:train_end,:]
x_train = np.vstack([x_sample[0:100,:],x_sample[300:500,:],x_sample[700:900,:],x_sample[1100:1150,:]])
y_train = np.vstack([y_sample[0:100,:],y_sample[300:500,:],y_sample[700:900,:],y_sample[1100:1150,:]])
print "x_train.shape = ",x_train.shape
sample_number = x_train.shape[0]
test_start = 1150
test_end = 1200
y_test = y_sample[test_start:test_end,:]
x_test = x_sample[test_start:test_end,:]
outdir = os.path.join(os.path.dirname(__file__),'stockClose/')
myNNmodel = MyNeurolNetworkModel()
# myNNmodel.outdir = outdir
myNNmodel.errorRate = 0.0105
myNNmodel.learningRate = 0.001
print "myNNmodel outdir = ",myNNmodel.outdir
if not os.path.exists(outdir):
os.mkdir(outdir)
# myNNmodel.train(x_train,y_train)
print "layerOne = ",myNNmodel.layerOne
print "myNNmodel train successfully ..."
y_test_predict = myNNmodel.predict(x_test)
from matplotlib import pyplot as plt
# x_train = x_sample[train_start:train_end,:]
x_train = np.vstack([x_sample[0:100,:],x_sample[300:500,:],x_sample[700:900,:],x_sample[1100:1150,:]])
y_train = np.vstack([y_sample[0:100,:],y_sample[300:500,:],y_sample[700:900,:],y_sample[1100:1150,:]])
print "x_train.shape = ",x_train.shape
sample_number = x_train.shape[0]
test_start = 1150
test_end = 1200
y_test = y_sample[test_start:test_end,:]
x_test = x_sample[test_start:test_end,:]
outdir = os.path.join(os.path.dirname(__file__),'stock_close/')
myNNmodel = MyNeurolNetworkModel()
# myNNmodel.outdir = outdir
myNNmodel.errorRate = 0.0105
myNNmodel.learningRate = 0.001
print "myNNmodel outdir = ",myNNmodel.outdir
if not os.path.exists(outdir):
os.mkdir(outdir)
# myNNmodel.train(x_train,y_train)
print "myNNmodel train successfully ..."
y_test_predict = myNNmodel.predict(x_test)
y_train = np.vstack([
y_sample[0:100, :], y_sample[300:500, :], y_sample[700:900, :],
y_sample[1100:1150, :]
])
print "x_train.shape = ", x_train.shape
sample_number = x_train.shape[0]
test_start = 1150
test_end = 1200
y_test = y_sample[test_start:test_end, :]
x_test = x_sample[test_start:test_end, :]
outdir = os.path.join(os.path.dirname(__file__), 'stock_close/')
myNNmodel = MyNeurolNetworkModel()
# myNNmodel.outdir = outdir
myNNmodel.errorRate = 0.0105
myNNmodel.learningRate = 0.001
print "myNNmodel outdir = ", myNNmodel.outdir
if not os.path.exists(outdir):
os.mkdir(outdir)
# myNNmodel.train(x_train,y_train)
print "myNNmodel train successfully ..."
y_test_predict = myNNmodel.predict(x_test)
from matplotlib import pyplot as plt
print "ysample = ",ysample.shape
# indexList = np.random.permutation(shape[0])
indexList = range(shape[0])
x_train = xsample[indexList[0:538]]
y_train = ysample[indexList[0:538]]
print "x_train.shape = ",x_train.shape
print "y_train.shape = ",y_train.shape
x_test = xsample[indexList[538:]]
y_test = ysample[indexList[538:]]
print "x_test.shape = ",x_test.shape
print "y_test.shape = ",y_test.shape
myNNmodel = MyNeurolNetworkModel()
myNNmodel.errorRate = 0.918
myNNmodel.layerOne = 15
myNNmodel.learningRate = 0.001
myNNmodel.trainTimes = 4000
# myNNmodel.batchSize = 20
y_predict = myNNmodel.predict(x_test)
print y_predict[0:10]
np.save('./npyfile/y_predict1',y_predict)
result = myNNmodel.f_measure(y_predict,y_test)
print "result = ",result
y_sample = np.zeros((shape[0]-related,1))
for i in xrange(shape[0] - related):
x_sample[i,:] = closeArray[i:i+related,0].reshape(1,related)
y_sample[i,0] = closeArray[i+related,0]
train_start = 600
train_end = 1150
y_train = y_sample[train_start:train_end,:]
x_train = x_sample[train_start:train_end,:]
test_start = 1000
test_end = 1200
y_test = y_sample[test_start:test_end,:]
x_test = x_sample[test_start:test_end,:]
test_start_1 = 400
test_end_1 = 600
x_test_1 = x_sample[test_start:test_end,:]
y_test_1 = y_sample[test_start:test_end,:]
mYnnModel = MyNeurolNetworkModel()
mYnnModel.errorRate = 0.040
if is_train == 1:
mYnnModel.train(x_train,y_train)
return crossDomainResponse({"code":200,"msg":"ok"})
else:
y_predict = mYnnModel.predict(x_test)
return crossDomainResponse({"code":200,"msg":"ok","y_test":y_test.tolist(),"y_predict":y_predict.tolist()})
print "ysample = ",ysample.shape
# indexList = np.random.permutation(shape[0])
indexList = range(shape[0])
x_train = xsample[indexList[0:538]]
y_train = ysample[indexList[0:538]]
print "x_train.shape = ",x_train.shape
print "y_train.shape = ",y_train.shape
x_test = xsample[indexList[538:]]
y_test = ysample[indexList[538:]]
print "x_test.shape = ",x_test.shape
print "y_test.shape = ",y_test.shape
myNNmodel = MyNeurolNetworkModel()
myNNmodel.errorRate = 0.918
myNNmodel.layerOne = 15
myNNmodel.isDropout = True
myNNmodel.learningRate = 0.001
myNNmodel.trainTimes = 4000
# myNNmodel.train(x_train,y_train)
'''
y_predict = myNNmodel.predict(x_train)
print y_predict[0:10]
for i in xrange(y_predict.shape[0]):
if y_predict[i] >= 0.5:
y_predict[i] = 1
columns = np.hsplit(dataset, 9) xsample = np.hstack(columns[0:8]) ysample = columns[8] shape = xsample.shape print "xsample = ", xsample.shape print "ysample = ", ysample.shape # indexList = np.random.permutation(shape[0]) indexList = range(shape[0]) x_train = xsample[indexList[0:538]] y_train = ysample[indexList[0:538]] print "x_train.shape = ", x_train.shape print "y_train.shape = ", y_train.shape x_test = xsample[indexList[538:]] y_test = ysample[indexList[538:]] print "x_test.shape = ", x_test.shape print "y_test.shape = ", y_test.shape myNNmodel = MyNeurolNetworkModel() myNNmodel.errorRate = 0.918 myNNmodel.layerOne = 15 myNNmodel.isDropout = True myNNmodel.learningRate = 0.001 myNNmodel.trainTimes = 7000 myNNmodel.train(x_train, y_train) print "train model successfully!"
columns = np.hsplit(dataset,9) xsample = np.hstack(columns[0:8]) ysample = columns[8] shape = xsample.shape print "xsample = ",xsample.shape print "ysample = ",ysample.shape # indexList = np.random.permutation(shape[0]) indexList = range(shape[0]) x_train = xsample[indexList[0:538]] y_train = ysample[indexList[0:538]] print "x_train.shape = ",x_train.shape print "y_train.shape = ",y_train.shape x_test = xsample[indexList[538:]] y_test = ysample[indexList[538:]] print "x_test.shape = ",x_test.shape print "y_test.shape = ",y_test.shape myNNmodel = MyNeurolNetworkModel() myNNmodel.errorRate = 0.918 myNNmodel.layerOne = 15 myNNmodel.isDropout = True myNNmodel.learningRate = 0.001 myNNmodel.trainTimes = 7000 myNNmodel.train(x_train,y_train) print "train model successfully!"
#!usr/bin/env/python
# -*- coding: utf-8 -*-
import numpy as np
import os
from myNeurolNetworkModel import MyNeurolNetworkModel
order = {"0": "Open", "1": "High", "2": "Low", "3": "Close", "4": "Volume*e-6"}
yahooData = np.load('yahoo_finance5.npy')
Open, High, Low, Close, Volume = np.hsplit(yahooData, 5)
shape = yahooData.shape
print("shape = ", shape)
print("-" * 80)
print("Close shape = ", Close.shape)
myNNmodel = MyNeurolNetworkModel()
kdj = myNNmodel.calculate_kdj(yahooData) #(None,3)
print("kdj = ", kdj.shape)
declose = myNNmodel.calculate_dclose(Close) #(None,4)
print("declose = ", declose.shape)
logfit = myNNmodel.calculate_logfit(Close) #(None,1)
print("logfit = ", logfit.shape)
closeExponent = myNNmodel.calculate_exponent(Close, exponent=0.9) #(None,1)
print("closeExponent = ", closeExponent.shape)
from sklearn.decomposition import PCA
pca = PCA(n_components=2)
newData = pca.fit_transform(np.hstack([Open, High, Low, Volume])) #(None,2)
import os
from myNeurolNetworkModel import MyNeurolNetworkModel
'''
每天的五个数据 high,low,close,open,adj_close,
'''
order = {"0": "Open", "1": "High", "2": "Low", "3": "Close", "4": "Volume*e-6"}
yahooData = np.load('yahoo_finance5.npy')
Open, High, Low, Close, Volume = np.hsplit(yahooData, 5)
shape = yahooData.shape
print "shape = ", shape
print "-" * 80
print "Close shape = ", Close.shape
myNNmodel = MyNeurolNetworkModel()
kdj = myNNmodel.calculate_kdj(yahooData)
print "kdj = ", kdj.shape
declose = myNNmodel.calculate_dclose(Close)
print "declose = ", declose.shape
logfit = myNNmodel.calculate_logfit(Close)
print "logfit = ", logfit.shape
closeExponent = myNNmodel.calculate_exponent(Close)
print "closeExponent = ", closeExponent.shape
# x_sample = np.hstack([Open,High,Low,Volume,kdj,logfit,closeExponent])
x_sample = np.hstack([Open, kdj, logfit, closeExponent])
y_sample = Close[1:]
train_start = 600
train_end = 1150
y_train = y_sample[train_start:train_end, :]
x_train = x_sample[train_start:train_end, :]
test_start = 1000
test_end = 1200
y_test = y_sample[test_start:test_end, :]
x_test = x_sample[test_start:test_end, :]
test_start_1 = 400
test_end_1 = 600
x_test_1 = x_sample[test_start:test_end, :]
y_test_1 = y_sample[test_start:test_end, :]
mYnnModel = MyNeurolNetworkModel()
mYnnModel.errorRate = 0.040
if is_train == 1:
mYnnModel.train(x_train, y_train)
return crossDomainResponse({"code": 200, "msg": "ok"})
else:
y_predict = mYnnModel.predict(x_test)
return crossDomainResponse({
"code": 200,
"msg": "ok",
"y_test": y_test.tolist(),
"y_predict": y_predict.tolist()
})
xsample = np.hstack(columns[0:8]) ysample = columns[8] shape = xsample.shape print "xsample = ",xsample.shape print "ysample = ",ysample.shape # indexList = np.random.permutation(shape[0]) indexList = range(shape[0]) x_train = xsample[indexList[0:538]] y_train = ysample[indexList[0:538]] print "x_train.shape = ",x_train.shape print "y_train.shape = ",y_train.shape x_test = xsample[indexList[538:]] y_test = ysample[indexList[538:]] print "x_test.shape = ",x_test.shape print "y_test.shape = ",y_test.shape myNNmodel = MyNeurolNetworkModel() myNNmodel.errorRate = 0.918 myNNmodel.layerOne = 15 myNNmodel.learningRate = 0.001 myNNmodel.trainTimes = 7000 # myNNmodel.batchSize = 20 myNNmodel.train(x_train,y_train,x_test,y_test) print "train model successfully!"
y_sample[1100:1250, :]
])
print "x_train.shape = ", x_train.shape
sample_number = x_train.shape[0]
test_start = 900
test_end = 1100
y_test = y_sample[test_start:test_end, :]
x_test = x_sample[test_start:test_end, :]
outdir = './images/'
if not os.path.exists(outdir):
os.mkdir(outdir)
myNNmodel = MyNeurolNetworkModel()
myNNmodel.inputNumber = 4
# myNNmodel.outdir = outdir
myNNmodel.errorRate = 0.01111
myNNmodel.learningRate = 0.001
# myNNmodel.train(x_train,y_train)
print "myNNmodel train successfully ..."
y_test_predict = myNNmodel.predict(x_test)
from matplotlib import pyplot as plt
plt.plot(y_test, 'ro')
plt.plot(y_test_predict, 'bo')
plt.plot(y_test, 'r-')
columns = np.hsplit(dataset, 9) xsample = np.hstack(columns[0:8]) ysample = columns[8] shape = xsample.shape print "xsample = ", xsample.shape print "ysample = ", ysample.shape # indexList = np.random.permutation(shape[0]) indexList = range(shape[0]) x_train = xsample[indexList[0:538]] y_train = ysample[indexList[0:538]] print "x_train.shape = ", x_train.shape print "y_train.shape = ", y_train.shape x_test = xsample[indexList[538:]] y_test = ysample[indexList[538:]] print "x_test.shape = ", x_test.shape print "y_test.shape = ", y_test.shape myNNmodel = MyNeurolNetworkModel() myNNmodel.errorRate = 0.918 myNNmodel.layerOne = 15 myNNmodel.learningRate = 0.001 myNNmodel.trainTimes = 7000 # myNNmodel.batchSize = 20 myNNmodel.train(x_train, y_train, x_test, y_test) print "train model successfully!"
#!usr/bin/env/python
# -*- coding: utf-8 -*-
import numpy as np
import os
from myNeurolNetworkModel import MyNeurolNetworkModel
order ={"0":"Open","1":"High","2":"Low","3":"Close","4":"Volume*e-6"}
yahooData = np.load('yahoo_finance5.npy')
Open,High,Low,Close,Volume = np.hsplit(yahooData,5)
shape = yahooData.shape
print("shape = ",shape)
print("-"*80)
print("Close shape = ",Close.shape)
myNNmodel = MyNeurolNetworkModel()
kdj = myNNmodel.calculate_kdj(yahooData) #(None,3)
print("kdj = ",kdj.shape)
declose = myNNmodel.calculate_dclose(Close) #(None,4)
print("declose = ",declose.shape)
logfit = myNNmodel.calculate_logfit(Close) #(None,1)
print("logfit = ",logfit.shape)
closeExponent = myNNmodel.calculate_exponent(Close,exponent = 0.9) #(None,1)
print("closeExponent = ",closeExponent.shape)
from sklearn.decomposition import PCA
pca = PCA(n_components = 2)
#!usr/bin/env/python
# -*- coding: utf-8 -*-
import numpy as np
import os
from myNeurolNetworkModel import MyNeurolNetworkModel
myNNmodel = MyNeurolNetworkModel()
v = myNNmodel.random_vector(10,100)
print("v = ",v.shape)
yahooData = np.load('yahoo_finance5.npy')
print("-"*100)
sample = yahooData[v]
print("sample.shape = ",sample.shape)