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 : Technical Analysis for Macintosh Users -- Ignore unavailable to you. Want to Upgrade?

To: xu, b. who wrote (850)	10/16/2001 9:22:02 PM
From: MechanicalMethod	Read Replies (1) \| Respond to of 1541

How are you using MAE+MFE? I'm thinking replace entryPrice with myFavorite signalLine. Then calculate AE + FE from mySignalLine every bar. Toss out 20% of the extreme bars so the remaining 80% equal the norm rather than outliers. Both minimum + maximum normal variance from the signal line. Project the signalLine 1 bar ahead with 80% AE+FE channels surrounding it. if signalLine=up then begin ..maxAE=.8(signalLine-L); ..maxFE=.8(H-signalLine); ..minAE=.2(signalLine-L); ..minFE=.2(H-signalLine); end; if signalLine=down then begin ..maxAE=.8(H-signalLine); ..maxFE=.8(signalLine-L); ..minAE=.2(H-signalLine); ..minFE=.2(signalLine-L); end; It's priced based rather than the bar based I want. I need to lookback(5) and drop the extreme so there's 4 bars left rather than what I did. Ratio's good but I'd prefer adjustable lookback... it's started... I've other stuff to do. I'd like to average it before projecting it, average(maxFE,3); Best regards, MM

To: xu, b. who wrote (850)	10/19/2001 10:04:01 AM
From: PuddleGlum	Read Replies (1) \| Respond to of 1541

I don't have any experience with PC emulators, but I think that Virtual PC may be the only one left on the market. There used to be another, but I don't remember the name of it. Regarding my backprop project... I haven't had time to work on it for a couple of weeks, but I didn't have to create a very large network for my Java backprop engine to become slow as molasses. My next step is to automate the training set generation, since I expect to try a wide variety of indicators in different training sets. One of the problems I see with minimizing the number of inputs to the network is that a chart has so much visual information that we use to make our decisions, and it's very hard to convert that visual info into a small number of inputs. pg