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 : Neural Nets - A tool for the 90's -- Ignore unavailable to you. Want to Upgrade?

To: Len Giammetta who wrote (65)	4/1/1998 1:46:00 AM
From: Jay Hartzok	Read Replies (1) \| Respond to of 871

Len, I've been running NeuroStock for the last several months and am getting fairly good results with it. You say you got lousy predictions. When I started using it, I got bad predictions too and finally determined that it was me making the errors, not the program. My errors involved not using enough comparison stocks, {and those that I did use were poorly selected} and not training enough hours {my best results come after 100 hours}. I won't even consider running a model now unless I have at least 8 comparison stocks in it, and I really prefer having 12 or more. All my comparison stocks are picked from the same business sector. I may have to go through 50 to 100 charts looking for stocks that seem to trade in tandem and in reverse to the target stock just to get 10 good choices. I also set up several different models on the same stock and test all of them to see which is the most accurate. Of course, doing all of this is very time consuming, but if you get an accurate model out of doing all of this, it is well worth the time. I've also found out, that for some reason, the Max Profit trading strategy produces the most accurate predictions. I have no idea why. It may be something that is unique to me or it may have something to do with the way I set up the model. Anyway, don't give up yet. Try making several new models on the same stocks that you got bad results on and vigorously train them and see if you can come up with an accurate one. Hope this helps. Good luck. Jay