Neural Nets - A tool for the 90's Message Board

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

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next

To: Optim who wrote (739)	4/5/2000 7:26:00 PM
From: xu, b.	Read Replies (1) of 871

Should we pick a single stock, a few stocks or an index or spyder? I vote for a set of stocks, just because I would want to determine fitness and eliminate spurious data by looking at the individual tickers rather than some masked reaction in an index or spyder, but thats just my vote This brings up another issue. In my testing, shorter periods of training data yeilds better profits, in sample and out of sample. I often worry about about overfitting the training data though. By using multiple stocks that have been screened to be relatively similar we can increase the amount of data out nets are exposed to, without increasing our training periods. Something that holds up across stocks will likely continue to hold up out of sample, and in realtime. Just my thought: Could we screen sectors with respect to SPX first? Then a second net will pick up strong stocks in the strong sectors. Spyder should be easier to predict compared to individual stocks or SPX. It's tradable. This is a bull model. Identification of trending vs. range trading is always important. I am very interesting in coming up with a valid shorting model or in another words, a bear model. Another important point is to make the system tradable. I know many systems are profitable on paper and practically impossible to trade. I vote for EOD models. Let me clarify what I said in my previous post. P/R should be the criteria for a good system. But it may not be necessary to use P/R as output of ANN to optimize. I have the impression that Profit may be able to do this sort of optimization. Tharp has a very good book on position sizing and expectancy. P/R is just another way of measuring expectancy of a system. In another words, how much you expect to make by risking 1 dollar. A number of trend following system has lower than 50% accuracy typically but they have positive expectancy because some big gains of small number of trades. Bai

Report TOU Violation

Share This Post

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next