Skip to content

"Making Predictions in Extremely Volatile Markets" - a project for CS229 Machine Learning at Stanford, Fall 2014

License

Notifications You must be signed in to change notification settings

acontry/cs229marketprediction

Repository files navigation

cs229marketprediction

My project "Making Predictions in Extremely Volatile Markets" for CS229 Machine Learning at Stanford, Fall 2014.

Key points

  • I predict the daily direction of movement in stock prices by utilizing classification machine learning techniques (SVM and logistic regression), training on 8 years of data and testing on the following 2 years.
  • This technique has a prediction accuracy of 52-60%, depending on the stock.
  • Even with this high error rate, it's easy to significantly boost risk-adjusted returns (measured by Sharpe ratio) by using the predictions in a simple trading strategy as compared to buy-and-hold. The trading simulation is quite lenient (no fees, slippage, etc) but the result still ain't bad!

About

"Making Predictions in Extremely Volatile Markets" - a project for CS229 Machine Learning at Stanford, Fall 2014

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages