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: xu, b. who wrote (750)	4/6/2000 7:21:00 PM
From: LastShadow	Read Replies (3) \| Respond to of 871

My personal opinion would be to take one telecom, one software, one computer PC, one bio, one semi, one ISP, one broadband, one internet - ecomm, one internet b2b, one portal This way we can pick the industry leader with best volume and price leadership, and others could refine into the concentration of choice. The purpose of taking awider spread of industries is that we will find that some tickers will work better on certain patterns/nmethods/strategies. If we concentrate on just a couple sectors, we are influeincing the direction the net will take. Also, in just a semi and bio selection, certain patterns will emerge and be reinforced by the sector news and leaders anyway. now, this isn't to say we can't take what we do and apply it to the stocks we wish to follow on outr own simultaneously, as it will provide some insight we could normally overlook. But to make a general net experiment, I think we should try to explore a diverse selection and see where are experiment leads us rather than start focusing too early. We carry predispositions into our design when we think we know where to look.

To: xu, b. who wrote (750)	4/6/2000 7:46:00 PM
From: xu, b.	Respond to of 871

oops, the matrix looks awful and I don't know how to fix it. HOpe you get what I want to say. As for the popup price condition, since many of the industry jumps are created by news events, there isn't a direct correlation between those occurances and preceeding price/volume action Let me come up with a wolf detector for discussion purpose. (I am not the inventor of this thing.) If you detect a significant increase in option activity, the wolf is coming. One market wizard got bit by that kind of wolves. You can make big bucks by following smart money's move. It even happened to me recently in equity market when I was trying to maintain a delta neutral (not exactly, with my own bias) position. Someone bought a stock from me at a price higher than day's high. Several days later, the stock ran up 7 points from 12. Of course, there was a company news that I did not know about. Someone knew that news! If they did anything, there must be a trace in price/volume. We should be able to find this kind of pattern or direct NN to look at these patterns. Lastshadow, I think I agree with you in that this kind of correlation is very difficult to discern and it comes down to a signal/noise issue that may not worth to study. Don't ask me how the guy bought the stock from me without showing the price in the day's high. I had a limit order in the book. If you have an answer, please let me know. I am still puzzllllling. Bai