I am continuously amazed by MSFTs tools. I have a model that I have developed that auto trades equities. I am in the process of writing out relevant statistics of performance etc to a database. It didn't take long to notice that certain equities are modeled better than others under certain circumstances, and that performance changes in complex ways from day to day, even intraday. What I do know to filter my universe of equities for that day is I use the neural net in my head to pick the equities that the model will be trading that day based on my experience with the model. However, I would like to make the selection completely scientific and automated. MSFT SQL Server has new algorithms to support this kind of data mining: http://msdn.microsoft.com/msdnmag/issues/04/09/AnalysisServices2005/default.aspx?loc=&side=true#a Nice! Does anyone have any experience with the data mining features of SQL server? nitro
Interesting stuff. I'd like to see how these data analysis tools could be applied towards either a stock selection model or a trading model.