Made some progress the weekend hunting down some raw data to work with. FINRA has daily files back to 2011. I'm experimenting with a few different ways of harvesting it. Ideally, I would rather go back to 2008 or earlier, so that I have a good data set to start with.
The first plan is to explore the question of whether or not it is possible to distinguish aggressive short selling from the regular sort. For aggressive short-selling, I intend to use a definition that is in line with the current regulation. Related to this is the question of whether or not it is possible to distinguish naked short selling in the current market.
Once I have a raw data set to play with, I want to experiment with applying a machine learning algorithm to it. In theory, computers can do a better job looking for trends then I could, and I m very curious to see what comes of this.
No comments:
Post a Comment