From theory to practice - page 562

 
Alexander_K:

Help.

In return, a ready-made Grail in VisSim+MT communication modules. All in all, a complete package, as promised.

It's not nice to promise something you don't have. And maybe you won't.) Don't give me the grail. (SIGHS) For free, that is.

So, what? Let's start with terms and definitions. What is a truth trend? Give me a definition.

 
Oleg Papkov:

Informatique isthe science of methods and processes for collecting, storing, processing, transmitting, analysing and evaluatinginformation using computer technology so that it can be used for decision-making[1]. From Wikipedia.

Even this list is quite independent fields and sciences.

 
Yuriy Asaulenko:

It's not nice to promise something you don't have. And maybe you won't have it.) No need for the Grail. (Laughs) For free, I mean for nothing.

So, what? Let's start with terms and definitions. What is a truth trend? Give me a definition.

(((( Again...) .... What's with the schoolyard rant?

All right.

As I understand it, a trend is a long non-trending price movement where there is a significant positive correlation ("memory") between the values in the price VR.

 
Alexander_K:

:(((( Again, it's .... What's with the schoolyard ranting habit?

Okay.

As I understand it, a trend is a long non-trending price movement where there is a significant positive correlation ("memory") between the values in the price VR.

Well, if we don't agree on the terms, we won't agree on anything).

What you have defined is exactly drift. Demolition and demolition, spit and forget. Let's say the ruble was demolished by an average of 4 kopecks a day over the year. For a speculator this is a flurry.

A trend is a sufficiently rapid and significant change of the price at the final interval of time. It is clear that "fast" and "significant" depend on specific applications.

 
Yuriy Asaulenko:

Well, if we don't agree on the terms, we won't agree on anything).

What you have defined is exactly demolition. Demolition and demolition, spit and forget. Let's say the ruble was demolished by an average of 4 kopecks a day. For a speculator this is a far-flung flat.

A trend is a relatively rapid and significant change in price on the definite time interval. It is clear, that "fast" and "significant" depend on particular applications.

Notorious so called spikes, which just break the usual diffusion equations. So?

 
Alexander_K:

The notorious so called jumps that just break the usual diffusion equations. So?

No, not jumps. Jumps are gaps.

So? So, at such motions, the LF components in the BP spectrum increase significantly. Remember the spectrum of 1(t) or the spectrum of the integral sin(x)/x, or go look it up on the net.

And since ACF, in normal families, is counted through BP spectra, no ACF will give absolutely nothing, until these very low-frequency components appear in the BP spectrum.

 
Yuriy Asaulenko:

No, not horse racing. Jumps are gaps.

Why? Because at such motions the LF components in the BP spectrum increase significantly. Remember the spectrum of 1(t) or the spectrum from the integral sin(x)/x, or look it up on the Internet.

And since ACF, in normal families, is calculated through BP spectra, no ACF will give absolutely nothing, until these very low-frequency components appear in the BP spectrum.

In general, you can't do without the Fourier transform. Did you really learn how to numerically express the phrase"significantly increase", did you make any tables?

 
Alexander_K:

In general, you can't do without the Fourier transform. Did you really learn how to numerically express the phrase"significantly increase", did you make any tables?

So, the problem is not in PF, but in detecting changes in the low-frequency part of the spectrum). And the low-frequency part appears not immediately, but as the pulse length increases. It cannot be detected immediately by any tricks neither PF, nor ACF, nor anything else. We can only analyze the difference in levels per dt and consider it a trend when the threshold is crossed. But this will only work as an additional method.

To detect low-frequency components you use low-pass filters - LPF. This is the most reliable way. The best of the simplest is the EMA, which has the lowest phase and group delay. The T parameter in the EMA is a guide and makes no real physical sense. If you do not have a correct LPF, you can only find the period of EMA, the movement of which corresponds to YOUR idea of the trend, experimentally on the graph. Then, we differentiate, which cuts off the zero harmonic and the very low-frequency components, and get a trend indicator. The level of the RNF corresponds (or rather, is proportional to) the current rate of trend. That's it.

 
Yuriy Asaulenko:

So, the problem is in the PF, but in detecting a change in the low-frequency part of the spectrum). Moreover, the low-frequency part does not appear immediately, but as the pulse length increases. It cannot be detected immediately by any tricks, neither PF, nor ACF, nor anything else. We can only analyze the difference in levels per dt and consider it a trend when the threshold is crossed. But this will only work as an additional method.

To detect low-frequency components you use low-pass filters - LPF. This is the most reliable way. The best of the simplest is the EMA, which has the lowest phase and group delay. The T parameter in the EMA is a guide and makes no real physical sense. If you do not have a correct LPF, you can only find the EMA, which movement corresponds to YOUR idea of the trend, experimentally on the graph. Next, we differentiate, which cuts off the zero harmonic and oscillating low-frequency components, and get a trend indicator. The level of the octa-zero corresponds to the current rate of the trend. That's it.

Thank you. I'll keep looking. Get your pockets ready, Yuri.

 
Alexander_K:


Hmmm, there you are... You're basically in all the forum threads at once and nowhere in particular

from the branch about machine learning, here are the incremental modules on Close draws the indicator that has a sliding window - where and to where? is it a regular MA?

Files: