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 (722)	3/28/2000 7:46:00 PM
From: xu, b.	Read Replies (2) \| Respond to of 871

I should thank Lastshadow for bringing me here! And I appreciate Optim's comments on NN development! I think trading x by watching y is a valid method. I traded Canada Trust's mutual funds this way several years back and I was not aware of Murphy's work on Intermarket analysis. You can call me stupid since I reinvent the wheel :>). A pattern I observed several years ago: small cap tends to rally a little bit later than blue chips. And I have dropped this pattern as trading setup because I was not able to construct a system out of this observation. Optim, is this pattern similar to what you mean about trading small cap by watching blue chip of same category? Bai

To: Optim who wrote (722)	3/28/2000 8:23:00 PM
From: xu, b.	Read Replies (1) \| Respond to of 871

Optim, Just a quick thought on intermarket analysis. If a scatter plot shows correlation between two sets of data (small cap vs. large cap for example), it may worthwhile to test it with NN. I am still a beginner in NN. I tried a combination of weekly data (MACD) and daily data (stoc, RSI, volume, RS vs. SPX...) to predict the percentage change of MA5 (and MA20) 5 days in the future and/or daily MACD. Not much luck so far. I tried 1 hidden layer, 2 hidden layers and three layers. The number of inputs varies from 5-11. Output layer has 1 or 2 nodes. Not much luck so far. It's funny that the highest contribution is from an input like C/MA20 which should be choppy compare to some smoothed indicators such as weekly MACD. From your post, I have a feeling that my prediction is too short in timeframe. Is there any NN in public domain that I can take a peek and get some inspiration? Brainmaker I ordered has an old user interface. Their book is an excellent intro one for me. I prefer Qnet2k so far. It offers an automatic normalization of inputs. I still plan to hold off my purchase of NS Trader or Profit until I really understand NN. Bai