def stockValueInFuture():
print(
"enter the code for the company you want to check the predictions of:")
companyCode = input()
print("Enter the No. of Days you want to predict the stocks for:")
n = int(input())
st = Stocker(companyCode)
st.create_prophet_model(n)
def listenToClient(client):
data = bytearray()
while True:
datachunk = client.recv(BUFFER)
if not datachunk:
break
data += datachunk
client.close()
data = data.decode().strip('\n').replace(" ", "").split(",")
requestID = data[0]
request_symbol = data[1]
days_into_future = data[2]
s3_client = boto3.client('s3')
obj = s3_client.get_object(
Bucket='elasticbeanstalk-us-west-1-643247086707',
Key=request_symbol + '.csv')
payload = str(obj['Body'].read()).replace(r'\n', '\n')
with open(request_symbol + ".csv", 'w') as f:
f.write(payload)
stock = Stocker(request_symbol, 'CSV', '')
predicted_price = stock.create_prophet_model(
days=int(days_into_future),
resample=False,
symbol=request_symbol,
requestID=requestID)
graph = open(request_symbol + "_" + requestID + ".png", "rb").read()
os.remove(request_symbol + '.csv')
r.set(request_symbol + "_" + requestID, graph)
os.remove(request_symbol + "_" + requestID + ".png")
r.set(request_symbol + "_price_" + requestID, predicted_price)
# In[38]:
f = open('msft6.txt', 'w')
f.write(
str(
microsoft.buy_and_hold(start_date='1999-01-05',
end_date='2002-01-03',
nshares=100)))
f.close()
# In[39]:
microsoft.changepoint_date_analysis()
# In[40]:
f = open('msft7.txt', 'w')
f.write(str(microsoft.changepoint_date_analysis()))
f.close()
# In[41]:
model, future = microsoft.create_prophet_model(days=30)
# In[42]:
f = open('msft8.txt', 'w')
f.write(str(microsoft.create_prophet_model(days=30)))
f.close()
from stocker import Stocker
hp = Stocker("HPQ")
stock_history = hp.stock
print(stock_history.head())
model, model_data = hp.create_prophet_model()
model.plot_components(model_data)
plt.show()
def GetStockPrediction(csvfile,ticker,inifile,litemode=False):
### MAKE HISTORY GRAPH.
hstg = 30 # number of days
# Define files and filetypes
ext = ".pdf"
ModZ =[str(stocksfolder+ticker+".HighLowClose" + ext), #0
str(stocksfolder+ticker+".ChangeVolume"+ext), #1
str(stocksfolder+ticker+".ChangePoint"+ext), #2
# str(stocksfolder+ticker+".ChangePointTrends"+ext),#3
str(stocksfolder+ticker+".evaluate_prediction"+ext),#4
str(stocksfolder+ticker+".TrendsA"+ext),#5
str(stocksfolder+ticker+".Naive" + str(hstg) + ext)]
#6
StockFiles=[str(stocksfolder+ticker+".HighLowClose" + ext), #0
str(stocksfolder+ticker+".ChangeVolume"+ext), #1
str(stocksfolder+ticker+".ChangePoint"+ext), #2
str(stocksfolder+ticker+".ChangePointTrends"+ext),#3
str(stocksfolder+ticker+".evaluate_prediction"+ext),#4
str(stocksfolder+ticker+".TrendsA"+ext),#5
str(stocksfolder+ticker+".Naive" + str(hstg) + ext),#6
# ACTUAL MACHINE LEARNING BEGINS
str(stocksfolder+ticker+".linear_regression_01"+ext)]#7
# print(csvfile)
# df = pd.read_csv(csvfile)
## START ANALSYS.
Months = str((today - datetime.timedelta(days=hstg)).date())
### Stocker is initialized and will retrieve the entire stock's data as it is read in the file.
stock = Stocker(ticker=ticker,
exchange='CSV',
csv_repository='stocks')
print(stock.stock.head())
plt.clf()
### plot_stock will plot the stock's (High, Low, Close) recent history. see 'hstg' for number of days
stock.plot_stock(start_date = str(Months),
stats = ['high', 'low', 'close'],
plot_type='hlc',
filename=StockFiles[0])
plt.clf()
### plot_stock will plot the Stock's (Daily Change, Volume) recent history. see 'hstg' for number of days
stock.plot_stock(start_date = str(Months) ,
stats = ['Daily Change', 'volume'],
plot_type='pct',
filename=StockFiles[1])
plt.clf()
stock.changepoint_date_analysis(filename=StockFiles[2])
plt.clf()
# stock.buy_and_hold(start_date=str(Months),nshares=1,filename=str(stocksfolder+ticker+".PredictedBuyHold"+ext)) # This function is broken.
# plt.clf()
stock.evaluate_prediction(start_date=str(Months),nshares=1,filename=StockFiles[4])
plt.clf()
## START BASIC PREDICTIONS BASED ON ANALYSIS
model, model_data = stock.create_prophet_model()
model.plot_components(model_data)
plt.savefig(StockFiles[5])
plt.clf()
stock.predict_future(days=hstg,filename=StockFiles[6])
plt.clf()
# START MORE ADVANCED PREDICTONS BASED ON DEEP LEARNING.
if litemode:
pass
else:
### changepoint_date_analysis is looking through historical data of the past 3 years to find the optimal time to buy or sell a stock.
print(inifile)
search_terms = inifile["STOCK PROFILE"]["search_terms"]
# print(search_terms)
bountylist = search_terms.split(",")
# print(bountylist)
stock.changepoint_date_analysis(search=bountylist,filename=StockFiles[3])
plt.clf()
# LINEAR REGRESSION 01
stock.Stocker_linear_regression(days=hstg, filename=StockFiles[7])
plt.clf()
## Merge files into a pdf.
if litemode:
ModZ = ModZ; print(ModZ)
else:
ModZ = StockFiles[:]; print(ModZ)
# thelimit = len(ModZ)
Merge_PDFS(ModZ,reportsfolder+ticker+"._REPORT.PDF")
from stocker import Stocker
amazon = Stocker('AMZN')
amazon.plot_stock()
# predict days into the future
model, model_data = amazon.create_prophet_model(days=90)
amazon.evaluate_prediction()
price = df.squeeze()
#print(price.head())
#print(price.index)
Target = Stocker(price, name=sid)
#Target.changepoint_prior_validation(start_date ='2015-12-03',end_date ='2018-12-21',changepoint_priors = [0.3,0.4,0.45,0.5,0.6])
#Target.plot_stock()
#原始參數預測
#model, model_data = Target.create_prophet_model(days=10)
#原始參數回測
#Target.evaluate_prediction()
#Target.changepoint_prior_analysis(changepoint_priors=[0.001, 0.05, 0.1, 0.2])
#Target.changepoint_prior_validation(start_date='2015-01-07', end_date='2018-12-21', changepoint_priors=[0.05,0.1,0.2,0.3,0.4])
#調整參數為0.4
Target.changepoint_prior_scale = 0.1
#Target.changepoint_prior_scale = 0.6
#修改後回測
Target.evaluate_prediction()
#修改後參數預測
Target.predict_future(days=30)
Target.predict_future(days=90)
model, model_data = Target.create_prophet_model(days=60)
#Target.evaluate_prediction(start_date='2017-12-15', end_date='2018-12-17', nshares=1000)
from matplotlib.pylab import rcParams
rcParams['figure.figsize'] = 20, 10
from sklearn.preprocessing import MinMaxScaler
from sklearn.linear_model import LinearRegression
from fastai.structured import add_datepart
import tensorflow as tf
from tensorflow.keras import layers
from sklearn import neighbors
from sklearn.model_selection import GridSearchCV
from pandas.util.testing import assert_frame_equal
goog = Stocker('GOOGL')
goog.plot_stock()
# Create model
model, model_data = goog.create_prophet_model(days=90)
goog.evaluate_prediction()
# Optimize the model
goog.changepoint_prior_analysis(changepoint_priors=[0.001, 0.05, 0.1, 0.2])
goog.changepoint_prior_validation(start_date='2016-01-04',
end_date='2017-01-03',
changepoint_priors=[0.001, 0.05, 0.1, 0.2])
# Evaluate the new model
goog.evaluate_prediction()
print(goog.evaluate_prediction(nshares=1000))
# Getting the dataframe of the data
goog_data = goog.make_df('2004-08-19', '2018-03-27')
print(goog_data.head(50))
from stocker import Stocker import matplotlib.pyplot as plt microsoft = Stocker(ticker='MSFT') stock_history = microsoft.stock # microsoft.plot_stock(start_date='2001-01-05', end_date='2002-02-05', stats=["Adj. Close"], plot_type="basic") # microsoft.buy_and_hold(start_date='2001-01-05', end_date='2002-02-05', nshares=200) model, model_data = microsoft.create_prophet_model() model.plot_components(model_data) plt.show() print(microsoft.weekly_seasonality) microsoft.weekly_seasonality = True print(microsoft.weekly_seasonality) model, model_data = microsoft.create_prophet_model(days=0) model.plot_components(model_data) plt.show() microsoft.weekly_seasonality = False microsoft.changepoint_date_analysis() model, future = microsoft.create_prophet_model(days=180)
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sat Sep 1 20:43:37 2018
@author: https://towardsdatascience.com/stock-prediction-in-python-b66555171a2
"""
#%%
from stocker import Stocker
teradyne = Stocker('TER')
#%%
model, model_data = teradyne.create_prophet_model(days=180)
#%%
teradyne.stock.head()
#%%
teradyne.evaluate_prediction(nshares=100)
#-------------
#%%
twtr = Stocker('TWTR')
#%%
model, model_data = twtr.create_prophet_model(days=200)
#%%
twtr.stock.tail()
'''
Created on 2019 M01 21
@author: kgap1
'''
from stocker import Stocker
aurora = Stocker('ACB.TO')
aurora.plot_stock()
# predict days into the future
model, model_data = aurora.create_prophet_model(days=90)
# import pystan
# model_code = 'parameters {real y;} model {y ~ normal(0,1);}'
# model = pystan.StanModel(model_code=model_code) # this will take a minute
# y = model.sampling(n_jobs=1).extract()['y']
# y.mean() # should be close to 0
f.close()
# In[18]:
britishmidland.changepoint_date_analysis()
# In[19]:
f = open('./BMI/bmi7.txt','w')
f.write(str(britishmidland.changepoint_date_analysis()))
f.close()
# In[20]:
model, future = britishmidland.create_prophet_model(days=30)
# In[21]:
f = open('./BMI/bmi8.txt','w')
f.write(str(britishmidland.create_prophet_model(days=30)))
f.close()
# Stock is an attribute of the microsoft object
stock_history = microsoft.stock
stock_history.head()
# A method (function) requires parentheses
microsoft.plot_stock()
microsoft.plot_stock(start_date='2000-01-03',
end_date='2018-01-16',
stats=['Daily Change', 'Adj. Volume'],
plot_type='pct')
microsoft.buy_and_hold(start_date='1986-03-13',
end_date='2018-01-16',
nshares=100)
microsoft.buy_and_hold(start_date='1999-01-05',
end_date='2002-01-03',
nshares=100)
model, model_data = microsoft.create_prophet_model()
# model and model_data are from previous method call
model.plot_components(model_data)
plt.show()
print(microsoft.weekly_seasonality)
microsoft.weekly_seasonality = True
print(microsoft.weekly_seasonality)
microsoft.changepoint_date_analysis()
def stockPrediction():
print(
"enter the code for the company you want to check the predictions of:")
companyCode = input()
st = Stocker(companyCode)
st.create_prophet_model()
StockName.predict(dates),
color='black',
label='Observations')
plt.plot(StockName.dates,
StockName.predict(dates),
color='black',
label='Observations')
plt.plot(StockName.dates,
StockName.predict(dates),
color='blue',
label='Predicted',
attr='bold')
plt.plot(StockName.dates,
StockName.predict(dates),
color='yellow',
label='Confidence Interval',
attr='block')
plt.plot(StockName.midpoint.random(dates),
label='Prediction Start',
color='red',
attr='dotted')
plt.xlabel('Date')
plt.ylabel('Price $')
plt.title(StockName, 'Model Evaluation from ', Startdate, 'to ',
Enddate)
model, model_data = StockName.create_prophet_model(days=no_of_days)
StockName.evaluate_prediction()
get_ipython().system('pip install quandl')
get_ipython().system('pip install fbprophet')
get_ipython().system('pip install pytrends')
from stocker import Stocker
import quandl
quandl.ApiConfig.api_key = 'AsJgrNdvc_MbnWxwNB4V'
import matplotlib.pyplot as plt
ms = Stocker('MSFT')
ms.plot_stock()
ms.changepoint_date_analysis()
ms.buy_and_hold(start_date='2017-09-01', end_date='2018-03-01', nshares=100)
model, future = ms.create_prophet_model(days=180)
