Out of sporting interest, I engaged in adaptive quote filtering - page 3
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
точно. еще добавлю, что время, в течение которого условно можно работать по "подогнанной" модели надо считать отдельно - видимо следует оценивать время корреляции по виду АКФ.
The application of the theorist is needed where it was invented for. That is, when it is lazy or difficult to calculate, and its application must be justified separately. I mean, that's why they usually stick it wherever they can. :)
But other than that, you're right.
Я там выше показал - что линейно взвешенная 3 ведет себя даже лучше чем ваш, ну запаздывание меньше.
Адаптивные фильтры это делается через спекрт, сначала детектится спектр потом по наибольшей мошьности отбирается три - тять частот и яильтруется .
Только скажу сразу результата особенного нет.
What you have described is not an adaptive filter. When building an adaptive filter, there is no need to define the spectrum. It also "selects three to five frequencies" according to some pre-determined proximity criteria - which is a separate problem.
The task of adaptive filtering itself was set, not least as a way to get rid of the spectral representation.
If one imagines an adaptive filter as
F(x,a), where
x is the input sequence
a is a filtering parameter
then constant adjustment of the filtering parameter a
and here already norm as closeness criterion would be quite natural and quite applicable.
The main thing is that it works :)
То, что Вы описали, не является адаптивным фильтром. При построении адаптивного фильтра нет необходимости в определении спектра. Кроме того, "отбирается три - тять частот" по каким-то наперёд заданным критериям близости - что представляет собой отдельную проблему.
Сама задача адаптивной фильтрации ставилась, не в последнюю очередь, как способ избавиться от спектрального представления.
Если представить адаптивный фильтр как
F(x,a), где
x - входная последовательность
a - парамаетр фильтрации
то постоянной подстройке подвергается параметр фильтрации а
и здесь уже вполне естественной будет и вполне применима норма в качестве критерия близости.
Главное, что это работает :)
Is that how it works !? -- :) I'm terribly happy for you - you won't believe it! -- keep using your approach to trading ! Are you sewing from a yacht now?
But so that you do not mislead others, you can mislead as you wish, it's still your own business ...
So that others don't go the other way, I'll explain.
I'll also give you a link to the MESA abaptive price series filtering.
And also a book, for future use - so that it would be clear - the trading task is somewhat different from the radar task. :))
2007 - A.A. Dakhnovich DISCRETE SYSTEMS AND DIGITAL SIGNAL PROCESSING
And the essence of adaptive filtering, although I'm not personally ( and don't want to be :) ) a big expert in this field. The essence is to assume that interference does not occur instantly, time to adapt "poor-quality" real filters (and they actually quality simply can not be 100% suppression) to current conditions. Adaptive filtering is based on the so-called "error", and its minimisation.
Here's a quick quote from that book.
An adaptive filter (AF) is a filter whose characteristics depend on the spectrum of the signal being processed. The main purpose of an AF is to improve reception or processing of information. An AF is a filter with variable coefficients.
AF design procedure consists of choosing a class of filter (FIR or IIR) and choosing the optimal algorithm for variable coefficient correction (adaptation).
AF consists of three elements:
1) a digital filter with variable coefficients;
2) error detection device;
3) the device implementing the adaptation algorithm.
In more complex cases, a different principle of adaptation, known as inverse adaptation, is used.
This adaptation process can be either one-cycle (one-step) or iterative. The main characteristics of an adaptation algorithm are the speed of convergence at a given error and complexity (amount of computation). The most commonly used algorithms are those based on the method of least squares (LOS). Depending on the characteristic of error averaging, global-adaptive and local-adaptive filters are distinguished.
Among the applications of AF, the main ones are:
1. Correction of distortions in the communication channel. In this case AF models the inverse characteristic of the communication system, so that the frequency response of the filter is inverse with respect to the frequency response of the communication channel.
2. Noise suppression. In this case, the AF is tuned by the pattern of interference, so that eventually subtract this interference from the received signal.
3. Compression of speech signals in systems with linear prediction (vocoders).
why so much sarcasm? :)
Yes, indeed, the characteristic depends on the spectrum of the signal being processed. It does, but it does not calculate spectrum (spectral response). It is unnecessary!
But... If you read to the end of definition, everything becomes clear :) - which elements AF is built of
It's also very well seen in the above diagrams.
Look into it more thoroughly
:)) I am not misleading anyone.
зачем же столько сарказма? :)
Да, действительно, характеристика зависит от спектра обрабатываемого сигнала. Зависит, но вычисление спектра (спектральной характеристики) при этом не производится. За ненадобностью!
Но... Если дочитать до конца определение, то всё прояснится :) - из каких элементов строится АФ
Это также очень хорошо видно на приведенных схемах.
Разберитесь более тщательно в этом вопросе
:)) А в заблуждение я никого не ввожу.
I don't think you get it - admit it... :)
That's why I'm being sarcastic, including the fact that I'm not an expert in this field. :)
What's it got to do with building or not building a spectrum? :) ... You notice you advised me to figure it out, haha... With adaptive filters the adaptation system can be ANYTHING, even based on email correspondence. :) Why did you come out with yours :) "the main thing is that it works" ... Without even understanding the essence of this TECHNICAL technique.
Well, since you don't understand it, let me explain - IT DOESN'T WORK. :)
Вы помоему так и непоняли - признайтесь...
Вот поэтому и сарказм, и в том числе про то что я не большой спец в этой области. :)
Причем тут построение или непостроение спектра - :) ... Вы заметьте посоветовали мне разобраться, у адаптивных фильтров система адапцации может быть любой даже и основанной на почтовой переписке. Зачем Вы вылезли с вашей :) "гловное это работает" ... не разобравшись даже в сути.
Ну раз вы не поняли обьясню - ЭТО НЕ РАБОТАЕТ. :)
Um... Admit it...
You ask: What does it have to do with building or not building a spectrum?
Isn't that your statement?
..................................................................................
SProgrammer wrote >>.
I showed there above - that the linear weighted 3 behaves even better than yours, well the lag is less.
Adaptive filters are done as a spectrum, first the spectrum is detected and then three to five frequencies are extracted with the highest power.
But I must say at once the result is not special.
And among other things, I did not get out, but gave my opinion.
And if you have "It doesn't work. :)", it doesn't mean it doesn't work for anyone :))