ml_for_stock_tranding.py

# -*- coding: utf-8 -*-
"""ML for Stock Tranding.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/15Ergxt4zGHY3Z9QWzXOtkZASfRD13btk

<h2> Machine Learning for Stocks Tranding </h2>

Step 1: Getting some historical data.

Libraries or Modules:

Yahoo Finance module.
Create Charts
Handle dataframes - Pandas
We need datetime functions
"""

import pandas as pd
import matplotlib.pyplot as plt
from datetime import date, time


#Install Yahoo Finance Libraries
try:
  import yfinance
except:
  !pip install -q yfinance
  import yfinance
try:
  import yahoofinancials
except:
  !pip install -q yahoofinancials
  import yahoofinancials

import yfinance as yf
from yahoofinancials import YahooFinancials

import warnings
warnings.filterwarnings("ignore")

#Download our Historical Data
df = yf.download('AAPL',
                 start = '2010-01-01',
                 end =date.today(),
                 progress=False,)

# Create a simple chart
df.plot(y='Close', title='APPLE STOCK')
plt.show()

"""Creating Technical Analysis Indicators and we are going to also plot them"""

try:
  import pandas_ta as ta #create basic technical analysis
except:
  !pip install -q pandas_ta
  import pandas_ta as ta

df['RSI(2)'] = ta.rsi(df['Close'],length = 2)
df['RSI(7)'] = ta.rsi(df['Close'],length = 7 )
df['RSI(14)'] = ta.rsi(df['Close'],length = 14)
df['CCI(30)'] = ta.cci(close= df['Close'],lenght = 30,high=df['High'],low=df['Low'])
df['CCI(50)'] = ta.cci(close= df['Close'],lenght = 30,high=df['High'],low=df['Low'])
df['CCI(100)'] = ta.cci(close= df['Close'],lenght = 30,high=df['High'],low=df['Low'])

#Drop NaN Values
df = df.dropna()

#Create a plot showing some of our indicators
df.plot(y='RSI(2)')
df.plot(y='CCI(100)')

#How's our current data frame going
df.head()

"""Prepare our data labelling

when we place our trade: Buy at Open (over the next day)

Formula:

Handling the returns this ways: Buy: Open(-1) - Closing: Open(-2)

We want to label our data in the following day

BUY SIGNAL: Open(-2) > Open (-1)

SEEL SIGNAL: Open(-2) < Open(-1)

Indicators the current situations, and our labels take a look ahead of time


"""

import numpy as np

df['LABEL'] = np.where(df['Open'].shift(-2).gt(df['Open'].shift(-1)),"1","0")
df = df.dropna()
df.head()

"""Creating a Neural Network

SkLearn
"""

import sklearn
from sklearn.neural_network import MLPClassifier
from sklearn.neural_network import MLPRegressor
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error
from math import sqrt
from sklearn.metrics import r2_score


#Fitting - Indicators and our Labelling

X = df[df.columns[6:-1]]
y = df['LABEL'].values

X_train , X_test, y_train, y_test = train_test_split(X,y,test_size = 0.3)

mlp = MLPClassifier(hidden_layer_sizes=(8,8,8),activation='relu',solver='adam',max_iter = 1000)
mlp.fit(X_train,y_train)

predict_train = mlp.predict(X_train)
predict_test = mlp.predict(X_test,)

from sklearn.metrics import classification_report,confusion_matrix

print('Train Data Accuracy')
print(classification_report(y_train,predict_train))

print( ' Testing Data Accuracy')
print(classification_report(y_test,predict_test))

"""Measure the returns obtained by trading our model"""

df['Prediction'] = np.append(predict_train,predict_test)
df['Strategy Returns'] = np.where( df['Prediction'].eq("1"),df['Open'].shift(-2)-df['Open'].shift(-1),0)

df['Strategy Returns'] = df['Strategy Returns'].cumsum()

df.plot(y='Strategy Returns')
plt.plot()

prediction = df.iloc[-1]['Prediction']

if prediction == "1":
  print("Today's return forecast: UP")
else:
  print("Today's retrun forecas: DOWN")