SI
SI
discoversearch

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

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: xu, b. who wrote (744)4/6/2000 7:45:00 AM
From: LastShadow  Read Replies (2) of 871
 
Yes, there are some problems with picking stocks from just one industry, such as semiconductor. There may be buying and selling patterns in semi's that are unique to that industry that are not evidenced in biotech. Also, if you attempt to go from an industry specific net to a stock specific net, you will find that the improvement in the net takes longer and requires further backtesting and more frequent retraining.

If the object is to show how to create a generic net and then finesse it to the specific secotr or stocks, then a more robust example would be to use a dozen stocks with one from eah major industry. The performance of the net will then tell you if your initial characteristic set and pattern discrimination method is better suited to one or another.

As for retraining, I retrained mine very frequently, but we could show the difference in performance by training daily, weekly, monthly, etc.

Equal dollar amount is fine

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 (I've looked). And while those events get a lot of press and make the mostuup lists, etc., they are not in the majority of stock movements. The reasonising is this. If we create a system that looks for strength, predictability and tradeability, it will catch those pops when they occur. More importantly, if those pops are gaps, then it is even better to model for that condition to determine entry and exit points.

The point for the investor/trader should be to optimize profit. Period. Whether the time frame is one hour or one year is initially immaterial. The development of a robust net should stive for consistency and accuracy rather than hunting volatility. Its really a philosophy of pace. I want every trade to make money, but if it doesn't, I want to exit with the minimal loss. I do not want to be trying to outguess the next flash in the pan, as that means I then have to look for intraday exits and entries.

Intray trading is a differnet animal, and more than a refinement of a position trading net - it is a whole different set of trading strategies and guidelines.

lastshadow
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext