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 : NeuroStock -- Ignore unavailable to you. Want to Upgrade?

To: CatLady who wrote (221)	10/12/1998 5:16:00 PM
From: Len Giammetta	Respond to of 805

CatLady, I train all my nets with back prop only. Per the documentation, some train better with back prop, some with annealing. I never really noticed any great improvement with annealing, and frankly I just don't have the patience. However, Jay's a big believer in annealing. Maybe he'll add some comments

To: CatLady who wrote (221)	10/14/1998 1:39:00 AM
From: Jay Hartzok	Read Replies (1) \| Respond to of 805

CatLady I also wonder how Simulated Annealing fits into the picture. Has anyone done any experiments with nets only trained on BackProp or only on Annealing? From the Help Files: Back Propagation - Enables Back Propagation training algorithm. Most useful during initial training. Simulated Annealing - Enables Simulated Annealing training algorithm. This very powerful training algorithm is used to fine-tune the neural network. It is very slow, sometimes taking several days to train a big network. I have never really questioned this {maybe I should}, so I have always used Back Prop for the initial training and fine tuned the net with Annealing. It seems to work well. Len has stated that he hasn't seen enough difference in the end result of his nets to warrant the extra training time that is necessary to complete annealing. I tend to disagree, as I have seen quite a bit of difference, but then I am providing Neuro with much more input data than he usually does. Since you have brought this up, I will take time over the weekend to do some tests. I'll post results next week. Jay