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: Optim who wrote (8)	12/14/1997 9:29:00 PM
From: Optim	Read Replies (1) \| Respond to of 871

Well I spent most of Sunday working on a new net in NS Trader. It is based mainly on momentum-type indicators, with some volume and indexes thrown in. The trading stats look good so far, with a 140% gain on average (over 15 securities, mainly tech stocks) for the period Nov 1/96 to Oct 31/97. I want to get the average error down a bit (from 3.71%) before I begin to test it in real-time though. I think I am going to purchase the GeneHunter package from Ward Systems to help find indicators that have good correlation with my target (5 day %change prediction). It is very time consuming to try the various combinations to get a decent net up and running, so this may give me an edge. I'll post some stats for the various systems I come up with so far in a future post. Maybe then we can get this thread moving a bit... Optim

To: Optim who wrote (8)	12/29/1997 12:54:00 PM
From: current trend	Read Replies (1) \| Respond to of 871

I'm still trying to understand how one would utilize neural nets in trading real money. Is this basically a black box out of which comes signals that may give a 55% edge? If so how do you integrate other forms of TA that might give conflicting recommendations though additive to the ultimate outcome? Of course psychologically how do you jump off a cliff if the system says to? If you selectively take the signals that are liked, results must suffer. How do you know when the NN no longer works without suffering an extreme drawdown, or if the drawdown is to be expected? Comments as to any other pitfalls that have been experienced through their use would be greatly appreciated. CT