Neuro-forecasting of financial series (based on one article) - page 5

 
f.t.:

I haven't read the article, I'm not (anymore) interested in NS, but

... ...you've got to fuck up.

______________________

And this is written by a programmer with experience...

You can go now, the thought is clear.

 
Once again I would like to point out that NS is a very good tool if you have a working trading system from the TA family. Just feeding the NS with all sorts of indicators, conversions etc. It won't help. NS will not pull the market. The NS can make your trading system better. This is true, but you should have the TS itself on the basis of technical analysis, some kind of..... And so, even if you fill it with inputs, it will still work poorly. Not durable......While trying to create an NS, always after training and a little bit of trading, the NS starts to lose... And picking up the parameters is not so easy.....
 
TheXpert:
And this is written by a programmer with experience...

and with experience, and not just programming ;)

I share it with him - he who needs it will take it as "someone's useful knowledge", and he who does not - it will seem like "trash" to those who do not.

who seeks what he finds....

Dixi

 

MS = Ph.D. by American standards

PhD = Doctor of Science. That's as high as it gets.

Google the internet for Sriram's research papers and find the following abridged version of his dissertation:

https://www.mql5.com/go?link=https://metapress.com/

Check out Sriram's LinkedIn profile:

https://www.mql5.com/go?link=http://www.linkedin.com/pub/sriram-lakshminarayanan/5/a53/968

Says he is currently working as a director at Mcube Investment Technologies. Runs Tactical Asset Allocation for pension funds and institutional investors. Checking out the company's website:

https://www.mql5.com/go?link=http://www.mcubeit.com/company.html

and read About Us:

The company is facilitating a movement to the new model of SMART Rebalancing® (Systematic Management of Assets using a Rule-based Technique), which represents a superior approach to managing asset allocation decisions within multi-asset portfolios.

I'm going to have a go at his method. Somehow this Sriram guy quickly traded his method for a mediocre job at Asset Allocation(https://en.wikipedia.org/wiki/Asset_allocation). But maybe he's using it there? Hardly, because the pension funds won't allow day trading with their assets. I knew some day-traders who worked for Asset Management Companies during the day and gambled on international exchanges at night.

 
gpwr:

I'm going to get into his method. Somehow this Sriram has quickly traded in his method for a mediocre job in Asset Allocation(https://en.wikipedia.org/wiki/Asset_allocation). Maybe he's using it there, though. I don't think so, because the pension funds won't allow day trading with their assets.


I am done with 1vr ))))...
 
gpwr:

MS = Ph.D. by American standards

PhD = Doctor of Science. There is no higher than that.


Vladimir, why are you in such a mess ...

A PhD is slightly lower than our PhD, while a Master's is simply a graduate who has defended his/her degree, as opposed to a Bachelor's who has not defended his/her degree :)

Did the cat survive?

 

The cat joo had. And he does.

 
Mathemat:

The cat joo had. And he does.


Thank you :)
 
gpwr:


I told you, by American standards. At one time I decided to get a PhD. I went to one of the American universities. So and so, "I have a diploma from the Moscow Institute .... (some people here know which one), with honours". The American bureaucrat pulls out a thick book from the shelf: "What institute did you get your degree from?" She finds my home institute among thousands of international institutions and says "Your degree is equivalent to a bachelor's degree". "Well, how come?" I exclaimed, "5.5 years of study, here are all my courses listed here, compare them with your bachelor's courses and their 4-year study". "Then he tells me that there are such offices in the U.S., which compare my courses for money and give me a paper certifying what American degree my diploma is equivalent to. "But you better not waste time and money on these offices, as our university is guided by this thick book and we do not care about all the other papers. That's how I, with gray hair, had to sit in a master's course with 20-year-old students. What can you do? PhD is easy. In the US, you can get it in two years after getting your master's, still with teenage pimples. In russia, to get a doctoral degree, you had to work for 20-30 years. And if a PhD was less than forty years old, then everyone looked at him as a genius. Although things have probably changed now.

There is a PhD and a doctoral degree. The closest analogues, respectively, are Doctor of Philosophy and Doctorate. Ph is a prefix that lowers the status :) Well, no one will forbid the student to attempt the doctorate, but it will not ensure a positive result. I am talking about American (international) standards.

Again: do not confuse the American title PhD with our doctor of science. They are incomparable.

 
nikelodeon:

Actually, it's overtraining. I'm surprised you don't know that. The conventional wisdom is that a network is over-trained when it starts working the way it did yesterday. That is, it does not emphasise key points in the input but starts producing the same signal as yesterday.....


I am also very surprised, how it is possible to have some opinion, having no grounds for it.
I already expressed my opinion about the term "overtrained" somewhere here. The term does not reflect the essence of the phenomenon at all. What English word is so translated? Rather, the term "rote" (from "crammed") or "learned" is more appropriate. The phenomenon is similar to that of a crammer or a crammer with a blank head, who understands nothing but has rote a paragraph word for word. The phenomenon occurs when teaching a volumetric network with a small number of samples. The network reacts properly to the training samples, but it is of no use, because it can accommodate the possibility of learning many more samples, i.e. it is just an empty head. The result it gives is whatever it is, not yesterday's. So, what you have such yesterday is not clear to me, some kind of miracles, some kind of magic overtraining.
It's like programming formulas, they say about memory leaks on every occasion. So when talking about networks - over-training, over-education, and few understand what it is.