Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : Galapagos Islands -- Ignore unavailable to you. Want to Upgrade?

To: Patrick Slevin who wrote (55717)	9/27/2006 7:31:56 PM
From: PuddleGlum	Read Replies (1) \| Respond to of 57110

I've got years of data (SPX back to 1950). I've got lotsa luck (the wrong kind). I've seen the flies. Where's the ointment? I've only been working on my own neural net solution for 3 years (with 150 pages of notes and test results merely on the foundational efforts), but I had a good teacher (Wasserman, author of a few books on the subject). He teaches better in person than he does as an author. You have to determine first how the experts use charts, then translate the data into a form that allows the network to see what the experts see. The big problem as I see it is that humans process a massive amount of data in the form of years per indicator with multiple indicators per chart page, and then look at additional charts and prior years of data. It's extremely difficult to train a network with such a huge volume of data at each snapshot, so there's one of the windows that lets the flies in. It'll probably be another couple of years before I convince myself that I can or cannot come up with a viable solution. Right now I feel like a gold miner with gold fever: "You're almost there, dig just a few feet deeper..." Then, "You've come this far already, try just a little harder." Siren songs, that what I hear.