Neural networks, machine learning, regressions, alogrithms based on bollinger bands, SMAs and momentum/contrarian strategies
https://github.com/kurumbus/algorithmic-trading/blob/master/interactive-brokers-sma.ipynb - a script which trades on Interactive Brokers using Simple Moving Averages. If you think it's too easy to be profitable - you're right.
https://github.com/kurumbus/algorithmic-trading/blob/master/momentum-contrarian-opt.ipynb - Attempt to optimize parameters for momentum/contrarian strategies.
Obviously if you play with numbers long enough, you can create an illusion of profitability. But median is obviously -10%
https://github.com/kurumbus/algorithmic-trading/blob/master/bollinger-bands.ipynb - Good old Bollinger Bands
The strategy produces net 1% in 6 years. Still better than your portfolio
https://github.com/kurumbus/algorithmic-trading/blob/master/DNNModel.py Training a sequential Keras Model
https://github.com/kurumbus/algorithmic-trading/blob/master/classification.ipynb - Simple Machine Learning Algorithm using scikit logistics regression
Produces 600% of profit on training set:
On real data... Not so much:
But it's regression, what do you want.
https://github.com/kurumbus/algorithmic-trading/blob/master/linear-regression-scikit.ipynb - Simple Linear Regression with scikit
Can previous day returns predict today's market? Nah, not really
https://github.com/kurumbus/algorithmic-trading/blob/master/neural-network-tensorflow.ipynb - a tensorflow neural network trying to use one-minute bars to find profitable opportunities.
Let's just say for now it can't find them.
https://github.com/kurumbus/algorithmic-trading/blob/master/neural-network-ibkr.ipynb - a working implementation of a neural network trading on IBKR (Interactive Brokers).