Machine learning in trading: theory, models, practice and algo-trading - page 1454
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Senior brothers in the trade, somewhere in the wilds of the branch suggested to test on SB, take the price instead of SB and see what will be acurasi and everything else, if it is clearly over 55% then obviously somewhere screwed, because SB can not predict much more than 50%, but with ZZ that price that SB equally "cool" predicted, this means what? That it is possible to trade on SB?Aliosha was getting 100% accuracy on the SB and didn't care much about it. It was just a seed for him.
I think he was just explaining the basics, here's the basics
No, I mean the picture.)
Tired of drooling, too beautiful.
It's in what time period? What is the secret technology?
training - 10 months, the test period after training - 6 months.
The technology is as simple as a felted tree: UGA and MLP with two inner layers.
training - 10 months, test period after training - 6 months.
The technology is as simple as a valenki: UGA and MLP with two inner layers.
What's UGA. Here https://en.wikipedia.org/wiki/UGA is the most appropriate Universal geometric algebra. This or something else?
Have you trieda forumsearch?)
1. If it is easy, give a concrete example with numbers, if you know how.
2. It is not necessary to advise ("take", "look") do yourself and prove by concrete examples your statement. And the reference to the "big brothers" ... You could have written simply: "I told a man here.
There are too many smart chatterboxes.
Aliosha was getting 100% accuracy in the SB and did not care much about it. It was nothing to him.
Just in case, I'll put here two datasets (lern and test for both separately) as an illustration, the first as features (f0...f9) macdac with different window and the target is also macdac (g0), and the second dataset chips also macdac, and the target direction ZZ.
On the first dataset it is difficult to pull higher than 51%, as it should be, it is logical, it is SB, the second dataset, without problems 65-70%, that on SB that on the price, miracles))))
I do not expect that anyone even once downloaded to check, because the intentions of most participants in a different plane, but since you asked, I provide an example.
The degree of detail in the process representation and the desired degree of predictability of that process are in inverse proportion. The only question is the margin of error. It is necessary to reduce the degree of detail as long as the error allows to identify the process (opportunities to make money from the process). Thus there may be processes (symbols) on which it is impossible in principle to make money (practice shows that not only may they be - they are) ...... Fortunately, as long as there are symbols available for the average trader on which it is possible to make money in the long term, Amen.
There are two contradictory, and at the same time logical, opinions... First: you have to choose as big and long history for testing (optimization), in order to describe as many possible possible outcomes that can happen supposedly in the future. Secondly, there is no sense in testing on a long history, since the market is constantly changing.
Do you feel the contradictions and ambiguity of both options? In the first case it will never be enough history, and in the second case it will never be enough to choose the minimum in a period (that period, which can be reliably taken as the WINDOW)
There are a lot of things there. For example - compound Poisson processes, which Alexander from TP branch invents and never invents)
Although R is the most nauseating language one could invent, the book and its topics are really good... I'm looking for examples in Python :)
https://nbviewer.jupyter.org/github/StuartGordonReid/Python-Notebooks/blob/master/Stochastic%20Process%20Algorithms.ipynb
Interesting topic, btw... I want to learn how to model jumps so I can subtract them from the model later
http://stuartreid.co.za/interactive-stochastic-processes/