Send the link below via email or IMCopy
Present to your audienceStart remote presentation
- Invited audience members will follow you as you navigate and present
- People invited to a presentation do not need a Prezi account
- This link expires 10 minutes after you close the presentation
- A maximum of 30 users can follow your presentation
- Learn more about this feature in our knowledge base article
Machine Learning with Python for Algorithmic Trading
Transcript of Machine Learning with Python for Algorithmic Trading
Trading Styles and Philosophies
Traditionally, there have been three major groups of traders/investors:
fundamental investors that analyze companies, their products, strategic position and financials in detail
technical investors that have a look at "charts" and analyze stock price movements
passive investors that do no analysis at all and invest in diversified portfolios
Prediction with Machine Learning
Let us use machine learning to "predict" a deterministic series of numbers.
Stock Market Prediction as a Classification Problem
Stock market prediction is a classification problem in that one is only/mainly interested in the direction of the market movement, not the size of the movement.
The question is: Will the market go up or down?
Logistic Regression for Classification
The big question is, are markets predictable after all?
For example, the efficient market hypothesis says that market prices "always reflect all available information", implying, for instance, that technical analysis (the study of charts) is pointless.
Another question is:
Can Machine Learning help with this problem? Or does it maybe contribute to problems?
Economist, 07. 10. 2016: Sterling takes a pounding
Calculating the log returns
Recently, a new type of trader/investor has risen to prominence.
Algorithmic traders make use of, among others:
mathematical, financial and data analytics algorithms
dedicated hardware approaches, like cloud, co-location, FPGAs, etc.
high performance approaches to trade in millisecond ranges
The holy grail of trading/investing is to predict in what direction the market will move over a horizon of
The best explanation seems to lie in the world of algorithmic trading—the computer programs that automatically generate transactions. Such programs have been blamed for “flash crashes” in the equity markets, particularly the May 2010 event which saw the Dow plunge nearly 1000 points in minutes before recovering.
The fitting step
We are only interested in the sign of the return ("up or down")
Applying logistic regression to the data set
Applying the logistic regression to BMW stock