Creating Datasets - I.py

## This script creates new normalized input variables and ML models to be used for prediction

import pandas as pd
import numpy as np
from sklearn import svm
from sklearn.model_selection import train_test_split
from imblearn.over_sampling import RandomOverSampler
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
from sklearn.metrics import precision_score
from sklearn.metrics import recall_score
from sklearn.externals import joblib
import datetime as DT


#Import Tickers (Symbols) file and VIX Data file
Tickers = pd.read_csv("C:\Users\lenovo\Desktop\Karishma\Stocks\SP500_Symbols.csv")
d0_vix = pd.read_csv("C:\Users\lenovo\Desktop\Karishma\Stocks\Scripts\RawData\VixData.csv")

#Create a list of parameters for numDays and minimum profit
numDaysList = [40]
minProfitList = [10]

#Create input variables for the model
for j in numDaysList:                                            
    for k in minProfitList:                                     
        for n in range(0,50):                                
            df =  pd.read_csv("C:\Users\lenovo\Desktop\Karishma\Stocks\Scripts\RawData\\"+Tickers.iloc[n]["Symbol"]+".csv")
            numDays = j
            minProfit = k
            
            #This date will be used to create the input variables and models
            Date = "2017-05-05"
            date_index=df[df['Date']==Date].index.tolist()
            df = df[:date_index[0]+1]
        
            for i in range(1,len(df)-numDays):
                #If the price of the stock increases by min profit or more within the next 'NumDays', then class = 1, else class=0
                df2 = df[i:i+numDays]
                maxPrice = df2['Adj. Close'].max(axis=0)
                Percent = 100*(maxPrice-df.ix[i]['Adj. Close'])/df.ix[i]['Adj. Close']
                if Percent >= minProfit:
                    df.set_value(i,'Class',1)
                else:
                    df.set_value(i,'Class',0)