Data mining

Dear Forum,

I have a three part query:

- Unfortunately my scripting is as good as non-existent, currently. So before embarking on steep learning curves, do people think it would be possible to build a Zorro script which mines data? The outline might be to genetically mine in-test data, then out-of-test data using all sorts of criteria.

- Although I've done several searches on the forum, I could find no reference to this being done before. Does anyone know if it has?

- Does anyone have any views on the validity of this process? I've read Dr Howard Bandy on this subject, amongst others, and the mining process seems to me to be useful, but the opinions of others would be welcome.

Thanks.

James.

Be careful of Selection bias and Data Mining bias
http://en.wikipedia.org/wiki/Data_dredging

Hi GPEngine

Thanks for that info. Always interesting to discover more information. I have looked at this, and the key seems to be statistically large enough data, good walk forward analysis, and a continual mining process, which allows older discoveries to die.

I'm hoping someone has attempted to do this in Zorro! As I say, my scripting is... well, basic! Therefore building on someone else's ideas would be good.