Has anyone tried analysing the performance of their stock portfolio using Decision Tree algorithms?

Decision Tree algorithms are available in open source packages like RapidMiner, and can be used to develop and test prediction models.For example, models of what kinds of stock are associated with what level of performance outcome, as defined by the presence/absence of a given set of attributes.

For example, the attributes of a stock might be whether they are included in various particular Stockopedia screens or not. The input needed is a matrix where rows = stocks held, columns = screen that may include a given stock and cell values = 1/0 whether a given stock is included within a given screen. Plus a final column where the outcome of interest is documented e.g. performance in the last six months (and is recoded into a nomimal format)

All the usual provisos apply:e.g. historical performance of a model is no guarantee of its future performance

Nevertheless, I would be intersted to hear if anyone has experiemented in this area, or is interested in doing so.

Unlock the rest of this article with a 14 day trial

Unlock with Email
Unlock with Google

Already have an account?
Login here