MachineLearningProject

Main program: BacktestingPortfolio.py

back_testing_portfolio(symbol, data_point, capital, initial_margin=5000, maint_margin=3500,

contract_size=1000, purchase_size=1)

symbol = 2 characters (Contract Symbol)

data_point to read from file

capital (initial capital)

initial_margin (initial margin of the contract)

maint_margin (maintenant margin of the contract)

contract size (multiply the price)

purchase_size (how many contracts purchased/sold per signal)

Folder Structure

Input File for Machine Learning: Users/..your_account../data/SmallHybrid/*.csv

Input Price: Users/..your_account../RA/MachineLearningProject/data/random_forest/csv/*.csv

Output Price: Users/..your_account../RA/MachineLearningProject/data/random_forest/*.csv

: Users/..your_account../RA/MachineLearningProject/data/random_forest/*.png

Logic

The software starts by checking the signal file generated by Machine Learning model in the ouptut location. If the signal files are not available there, the software will generate the file by calling random_forest module. For each contract, the software will run 10 rounds by rolling over every 1000 data points. The software will generate P&L and price plots for each round and also .csv files that contain raw data and also the training reports.