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 : NeuroStock
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: CVDave who wrote (735)4/9/1999 12:34:00 AM
From: Jay Hartzok   of 805
 
Dave,

The one problem is the one that was posted at the Yahoo club, which involves the net using data in the files that is before the training start date. It slows down the training process and if you edit the data files as Mark has suggested, and are using any filter inputs, the first 50 calculations will be wrong and these bad calculations seem to effect the rest of the net. Even with no filter inputs the first calculations will be wrong for the length of the input's influence period.

If you allow enough data to exist in front of the training start date to produce the proper calculations, then it appears to me that after the net is trained and you start adding data, that the net stops learning and will repeat the same type of errors over and over again. If you retrain the entire identical net from the beginning the errors will not be there at all. These results that I've seen are very confusing to me and destroy my faith in the nets. Something just doesn't seem right to me and that's why I stopped using it for the time being.

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

HomeMailHotSubjectMarksPeopleMarksKeepersSettings
Terms Of UseContact UsCopyright/IP PolicyPrivacy PolicyAbout UsFAQAdvertise on SI
© 2025 Knight Sac Media.  Data provided by Twelve Data, Alpha Vantage, and CityFALCON News