Search | User Listing Vidsite | Calendars | Quotes
Home Page ->  General Discussion -> Miscellaneous Discussions -> View ThreadLogon (or Register, or Join TradeTight)

You are logged in as a limited-access Guest.To join TradeTight, first read the info in the Organization & Content room, then click the link above. 

James 1:5 (NIV) ... If any of you lacks wisdom, he should ask God, who gives generously to all without finding fault, and it will be given to him.


Basics of Neural Networks
Jump to page : 1
Now viewing page 1 [50 msgs/pg]
Jump to room :
RossKovacs
Posted 12/15/2012 9:55 AM (#4218)
Subject: Basics of Neural Networks



Veteran

Posts: 131
10025
Location:
USA: GA, Snellville
Hello,

The link below contains a recent article on Neural Networks, their history in trading, and recent developments.
The author (Murray A. Ruggiero, Jr.) has been around a long time and has authored a couple of books (check Amazon for examples).
The article is interesting to me, and I am looking for other viewpoints on the article:
http://www.futuresmagdigital.com/futuresmag/201212/?pg=54&pm=1&u1=f...

Top of the page Bottom of the page
JimDean
Posted 12/15/2012 10:22 AM (#4219 - in reply to #4218)
Subject: Basics of Neural Networks



Owner/Admin

Posts: 3925
2000100050010010010010025
Location:
USA: GA, Lawrenceville
Very general discussion - interesting historical recap.

Major point seems to be that NN's by themselves can't be used to create good signals ... rather, they are best used to improve signals of existing systems. Not sure if true, but worth considering!

That's what Ed taught us in the initial (and only so far) AI course in OTU.

Followup articles should be more useful imho.
Top of the page Bottom of the page
RossKovacs
Posted 12/15/2012 10:40 AM (#4220 - in reply to #4219)
Subject: Basics of Neural Networks



Veteran

Posts: 131
10025
Location:
USA: GA, Snellville
I agree that a major point was "NN's by themselves can't be used to create good signals ... "
Which makes me wonder about the appropriate structure for a Strategy.
For example, NSP-41 has two blocks: NSP-41 and Trade Plan.
Question: what is the appropriate structure for a Strategy that uses the ARM?

The author also emphasized data pre-processing.
He gave Mark Jurik's "Level 0" and "Level 1" method as a simple way to avoid the extensive data pre-processing that is ordinarily needed.
Question: do we know what (if any) data pre-processing is done by the ARM?
Top of the page Bottom of the page
JimDean
Posted 12/15/2012 11:02 AM (#4221 - in reply to #4220)
Subject: Basics of Neural Networks



Owner/Admin

Posts: 3925
2000100050010010010010025
Location:
USA: GA, Lawrenceville
ARM discussions need to be held in the NClub Room (click here), due to confidentiality requirements.

This thread is continued in that room, here:
http://tradetight.org/forums/thread-view.asp?tid=833
Top of the page Bottom of the page
RossKovacs
Posted 1/26/2013 9:21 AM (#4331 - in reply to #4218)
Subject: Basics of Neural Networks



Veteran

Posts: 131
10025
Location:
USA: GA, Snellville
The link below contains the second article on Neural Networks, their history in trading, and it starts to get into some significant topics.
The author (Murray A. Ruggiero, Jr.) has been around a long time and has authored a couple of books (check Amazon for examples).
The article is interesting to me, and I am looking for other viewpoints on the article (no discussion of ARM, per Jim's post):

http://www.futuresmagdigital.com/futuresmag/201302/?pg=50&pm=1&u1=f...


Top of the page Bottom of the page
JimDean
Posted 1/26/2013 9:59 AM (#4332 - in reply to #4331)
Subject: Basics of Neural Networks



Owner/Admin

Posts: 3925
2000100050010010010010025
Location:
USA: GA, Lawrenceville
This article is much more interesting to me than the prior one. Although I'm not enough of a math guru to totally follow all his comments, he is a good enough writer to figuratively describe his points along the way, in parallel with his more precise formulations.

The "flavor" and "tone" of the article seem to imply that this methodology is a work in progress rather than a functional engine - it seems to be a "scope of effort" for future work.

As is often the case when I read stuff like this, even if I don't fully follow what's being said (or even if I don't agree with some things I do understand - not the case here) … there usually are four or five ideas that pop into my head in relation to other methods or applications that I can utilize immediately.

For example, his discussions of preprocessing in combo with his hi-falutin' hyper-plane logic got me to thinking of some cruder but useful ways to scan lists for symbols that would offer good RTM trading opportunities.

That kind of thing alone makes articles like this valuable to me. Thanks for posting it, Ross!
Top of the page Bottom of the page
RossKovacs
Posted 4/8/2013 7:47 AM (#4467 - in reply to #4218)
Subject: Basics of Neural Networks



Veteran

Posts: 131
10025
Location:
USA: GA, Snellville
The link below contains the third article on Neural Networks.
The author (Murray A. Ruggiero, Jr.) has been around a long time and has authored a couple of books (check Amazon for examples).
The article is interesting to me, and I am looking for other viewpoints on the article (no discussion of ARM, per Jim's post):

http://www.futuresmag.com/2013/03/26/casting-neural-nets-into-moder...


P.S. If you enjoy these free articles, you can enter the name Ruggiero in the search box in the upper right corner and find many other articles by him. He has written some good stuff and I believe he is a contributing editor to this free magazine.

Edited by RossKovacs 4/8/2013 8:13 AM
Top of the page Bottom of the page
RossKovacs
Posted 12/20/2013 1:42 PM (#5507 - in reply to #4218)
Subject: Basics of Neural Networks



Veteran

Posts: 131
10025
Location:
USA: GA, Snellville
The link below contains a fourth article on Neural Networks.
The author (Murray A. Ruggiero, Jr.) has been around a long time and has authored a couple of books (check Amazon for examples).
The "Support vector machine (SVM) models" in this fourth article were introduced in the third article above, but I do not think that they can be applied in the ARM.
The article is interesting to me, and I am looking for other viewpoints on the article (no discussion of ARM, per Jim's post):

http://www.futuresmag.com/2013/08/01/advanced-model-building-for-tr...

Top of the page Bottom of the page
Jump to page : 1
Now viewing page 1 [50 msgs/pg]
( E-mail a Link | Printer Version | Search Room )

Owner of site: Jim Dean -- Forum content is confidential, and may not be distributed without written permission.