Adaptive digital filters - page 25
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
вот и получается чтобы проверять идеи программисту надо n дней а мне 1000n (это уж я утрирую) так как делаю все представления в уме и думать об этом приходится пастоянно либо со временем мысли рассеиваются либо вообще перескакнуть стараются на что то другое потом стараешся вернутся а часть мыслей уже ушла и ...... да и комп сволоч тормозит (
Thoughts are easy to write down on paper.
1. As you write, you formulate it more clearly; if you cannot formulate it, there is no thought, but an illusion.
2. Genius ideas are not forgotten so quickly. After a while it's useful to look back and... laugh
3. A written down idea is easier to translate into an algorithm and therefore into code. Especially if you don't have programming experience.
4. In the process of writing it becomes obvious that 9 out of 10 ideas are untenable. Accordingly, less time is spent on manual checking or coding.
There is something to the idea of adaptability, though. Suppose we adapt the TS to an existing BP area. Due to the non-stationarity of the market the next interval is likely to be different from the previous one and the TS may become less profitable or even unprofitable. I observe that customized systems (by the optimizer) do not change their profitable characteristics abruptly, but gradually lose their profitability up to unprofitability. It is natural, they have caught trend parameters, adjusted, and then the trend finishes and another one starts. The task of adaptation is to follow the change of non-stationary market step by step in the hope that there will be a profitable section ahead.
Это не совсем корректно. Да, сигналы могут быть любыми, но в основе применимости того или иного метода фильтрации лежит предположение о вполне определенной модели полезного сигнала. В случае шумового (случайного) сигнала-тоже.
Yes everything is quite correct. Where did I write about "forbidding" the acceptance of any particular signal model?
Side note: you need a model to take the spectrum, but you don't need a model to register any signal at all.
Do you have experience with Hilbert-Huang transformations?
У Вас есть опыт работы с преобразованиями Hilbert-Huang'а?
I haven't gotten to it yet, although I've left a note for myself. I think it's an interesting direction.
еще не добрался, хотя пометку для себя оставил. Мне кажется, это интересное направление.
I would be happy to cooperate with you in this regard.
Please get in touch when you get your hands on it.
Походу уже команда собирается:)))
I'm thinking about it too. I have a similar idea for decomposing kotir into its components, if it fails, I'll do a huang. Did you find the books?
Here it is (I haven't looked more closely yet) http://prodav.exponenta.ru/mapsite.htm
Буду рад сотрудничать с Вами в этом направлении.
Обращайтесь, когда руки дойдут.
Joint research can bring good results and save time. Because I believe that a "complex system" (such as a market) can be understood by a comparable system of "complexity". And this "learning phenomenon" system with the right qualities can theoretically self-organise from other systems. But that's just so, philosophy in the sense of a joke. :о))))))
There are 2 subtleties:
(1) I will still be 2-3 months very busy
(2) not a mathematician and not a DSP professional