Machine learning in trading: theory, models, practice and algo-trading - page 806

 
Maxim Dmitrievsky:

https://towardsdatascience.com/validation-methods-for-trading-strategy-development-1efea8284b02

I ditaph... everything, but I only throw in what I like or agree with, or have written about it myself before.

I wrote about it and cited an excerpt from the article to prove my point

If you doubt my cognitive abilities :D

I saw a denial of the usefulness of the available method, but you didn't provide an alternative solution. That's why I ask for the link, to see if the solution/replacement is offered in the full text.
 

Now, the article says "out-of-sample" in black and white.


And you've added a few more concepts to it.

Maxim Dmitrievsky:

For fans of crossvalidation, test sampling, OOS and other wild things, I will not tire of repeating:

I understand that for you it's all the same thing, so you've lumped it all together. But no, they are different things, there are several kinds of crossvalidation alone.

 
elibrarius:
I saw a denial of the usefulness of the available method, but you didn't provide an alternative solution. That's why I ask for the link, to see if a solution/replacement is proposed in the full text.

no solution is offered because no one knows it

But the points of inapplicability of some methods for non-stationary problems are emphasized

 
Dr. Trader:

Now the article says "out-of-sample" in black and white.


And you've added a few other concepts to it.

I understand that for you it's all the same thing, so you've lumped it all together. However, no, these are different things, there are several types of crossvalidation alone.

To me and to all normal people it's all the same thing - a fit

which is inapplicable to non-stationary problems.

read the manuals.

 

Well, don't speak for others......

I was right about you after all :)

 
I, for example, have found a way to evaluate without using the feedback loop. That's what I've been telling you lately. And this method is GUARANTEED to ensure the good work of the TC on the OOS. But where to me with my nonsense, as some people express here. And with regression you seem to have screwed up, because you really missed a chance to get a method of model preparation. Exactly the METHOD, which could be applied to your data and your TS....
 
Maxim Dmitrievsky:

no solution is offered because no one knows it

but the points of inapplicability of some methods for non-stationary problems are emphasized

So you're screwed. You're a cardboard fool. I know.... Come to think of it....

 
Dr. Trader:

Well, don't speak for others......

I guessed about you, though :)

are you talking some conspiracy bullshit that you guessed?
 
Mihail Marchukajtes:

So you're screwed. You cardboard fool. I know.... Come to think of it....

Do not bury your knowledge in 10 pages of this forum again, please

write meaningfully ))))))

 

And I think a lot of people here know that. Those who look at things realistically, without rose-colored glasses.

I need a period of testing and validation to assess the adequacy of the model, the OOS only a little bit, in order to be finally convinced. But there is still enough of it to work on it...

I wrote about it some time ago, but all by ear...