Forums  > General  > Deep Learning and Technical Analysis  
     
Page 1 of 1
Display using:  

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2017-09-19 01:53
It was only a matter of time.

Anyone want to write a paper on deep learning and phrenology?

I am calling #peakdeep

jslade


Total Posts: 1168
Joined: Feb 2007
 
Posted: 2017-09-19 07:03
This has been going on for decades. They just didn't call it dweeb learning back then.

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.134.1114&rep=rep1&type=pdf

http://ieeexplore.ieee.org/document/5726498/

https://people.eecs.berkeley.edu/~akar/IITK_website/EE671/report_stock.pdf

I know an honest academic working on this who achieved a small, publishable success; I suggested he try lars with the same inputs. Guess what happened.

"Learning, n. The kind of ignorance distinguishing the studious."

Baikinman


Total Posts: 23
Joined: Aug 2010
 
Posted: 2017-10-17 11:17
Isn't this the same kind of pattern recognition which is done e.g. by fitting the equation of D. Sornette (an exponential change with an added sin with variable frequency and amplitude)? I started to play with AI a couple of months ago, and this is my feeling at the moment, but I am stil a beginner iin this field. Any experiences?

Maggette


Total Posts: 1097
Joined: Jun 2007
 
Posted: 2017-10-17 17:11
I don't see any connection between the two approaches.

Maybe both are considered dangerous or pointless because of the danger of overfitting

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

katastrofa


Total Posts: 458
Joined: Jul 2008
 
Posted: 2018-03-02 01:10
Pentagon's buying it: https://thebulletin.org/project-maven-brings-ai-fight-against-isis11374

eye51


Total Posts: 200
Joined: Oct 2004
 
Posted: 2018-03-05 11:58
@katastrofa,

I am not sure I understand. I can see clear advantages / applications of AI for defence.

Peace means reloading your guns

katastrofa


Total Posts: 458
Joined: Jul 2008
 
Posted: 2018-03-06 09:56
Exactly.

eye51


Total Posts: 200
Joined: Oct 2004
 
Posted: 2018-03-07 12:36
Ah :)

Peace means reloading your guns

steevo


Total Posts: 9
Joined: Feb 2019
 
Posted: 2019-02-11 23:26
@jslade what is "lars" and what happened?

in reference to
"This has been going on for decades. They just didn't call it dweeb learning back then.

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.134.1114&rep=rep1&type=pdf

http://ieeexplore.ieee.org/document/5726498/

https://people.eecs.berkeley.edu/~akar/IITK_website/EE671/report_stock.pdf

I know an honest academic working on this who achieved a small, publishable success; I suggested he try lars with the same inputs. Guess what happened."

Maggette


Total Posts: 1097
Joined: Jun 2007
 
Posted: 2019-02-12 07:27
My guess is he is talking about least angle regression. A little bit smarter way than stepwise regression for a large set of features

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

jslade


Total Posts: 1168
Joined: Feb 2007
 
Posted: 2019-02-12 20:22
The LARS results were better than deep learning results on the same feature set for reasons which should be obvious to anyone reading this forum.

"Learning, n. The kind of ignorance distinguishing the studious."
Previous Thread :: Next Thread 
Page 1 of 1