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: mark_andrews who wrote (826)	9/8/2000 11:37:40 PM
From: F Robert Simms	Read Replies (1) \| Respond to of 871

That is a great post to Larry! I agree with your points. Your post shows another point why a software developer might not be interested in trading. It just takes an awful lot of work to find good inputs. I think I found a great input, but I guess it will take your six months to find out how good it really is.

To: mark_andrews who wrote (826)	9/11/2000 8:28:31 PM
From: Larry Livingston	Read Replies (1) \| Respond to of 871

Hey Mark, I've noticed the same thing. Getting a model that has performed well on out of sample data is no guarantee of good results. In fact if your data only covers the last five or six years you get a natural upward bias in your model which will give skewed results that actually seem great when applied to stocks that have performed well in the bull market. I'm curious to hear how you've automated. I've wanted to do that so that I could experiment with training thousands of different nets using different combinations of inputs. The only way to do it that I found would be to buy an neural engine as ActiveX component and then write my own code using a genetic algorithm. But perhaps you've discovered a simpler method. All the best, Larry