Machine learning in trading: theory, models, practice and algo-trading - page 259
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
mytarmailS:
Guys, who'll help me with this question http://ru.stackoverflow.com/questions/612114/%D0%BD%D1%83%D0%B6%D0%B5%D0%BD-dtw-%D0%B0%D0%BB%D0%B3%D0%BE%D1%80%D0%B8%D1%82%D0%BC-%D0%B0%D0%BD%D0%B0%D0%BB%D0%BE%D0%B3%D0%B8%D1%87%D0%B5%D0%BD-%D0%B0%D0%BB%D0%B3%D0%BE%D1%80%D0%B8%D1%82%D0%BC%D1%83-%D0%B8%D0%B7-dtw-package, I'll show how I can really predict the market, without water and exaggerations within reasonable limits of course
=====================================
Check out the package "dtwclust". This is a newer package and, as they say in the description, with new methods of calculating distances. All functions are in C. I didn't use it and didn't understand the details.
Where do you see the uses for it?
Vladimir Perervenko:
Look at the package "dtwclust". This is a newer package and, as they write in the description, with new methods of distance calculation. All functions are written in C. I haven't used it and haven't gone into details.
I've already tried everything, including dtwclust, for several weeks. I need exactly that algorithm, I don't know why similar algorithms are not applicable, even the same algorithm with a slightly different method of distance calculation works much worse
Vladimir Perervenko:
Where do you see the application for this?
I've created some know-how which can predict trends, or rather trend reversals, and it is realized with the help of the dtw algorithm, but dtw is just a tool, it won't work on its own.
I also want to note that there are no parameters in the method
Help me to implement, or rather speed up the algorithm, and I will tell you everything and show you
For several weeks I have already reviewed and retested everything possible, including dtwclust, I need exactly that algorithm, I do not know why similar algorithms were not applicable, even the same algorithm but with a slightly different method of distance calculation works orders of magnitude worse
I have created some know-how that can predict trends, or rather reversals of trends, and it is realized with the help of the dtw algorithm, but dtw is only a tool, it won't work on its own - it uses forecasting directly as it is
I also want to note that there are no parameters in the method
Help me implement, or rather speed up the algorithm and I will tell you everything and show you
Will it work?
Will it work?
Of course not, I'm telling you, you need exactly what you need, not everything with the prefix dtw
So there's, like, an algorithm from the wiki.
The algorithm from the wiki counts the distance by "symetric1" and I need "symetric2"
I don't know how "symetric2" is counted...
That's my question on stackowerflow above
Gentlemen, what if we improvise something like numerai but with real data and compete?
I propose to discuss conditions. For example, we will use a moderate high-frequency in the Russian market, not hard MM, take the flow of prices, volumes, open interest for local liquid futures and prices of major currency pairs of forex and a set of liquid Western indices, futures, etc. Synchronized seconds prices. The data is publicly available.
Forecasting our futures. The holding horizon is minutes, but it depends on who wants, the thing is that there will be not much data, for example, a week and it will not be possible to train trading hours/ days.
The last day of the week is not known, we need to teach the system to trade it profitably.
Gentlemen, what if we improvise something like numerai only with real data and compete?
I propose to discuss conditions. For example, we will use a moderate high-frequency in the Russian market, not hard MM, take the flow of prices, volumes, open interest for local liquid futures and prices of major currency pairs of forex and a set of liquid Western indices, futures, etc. Synchronized seconds prices. The data is publicly available.
Forecasting our futures. The holding horizon is minutes, but it depends on who wants, the thing is that there will be not much data, for example, a week and it will not be possible to train trading days.
From a week of data, the last day is not known, we need to teach the system to trade it profitably.
It will be very difficult to determine the winner. Someone will martingale and win with money and not with accuracy of the model and argue about winning with someone who seems to win with accuracy. Someone will just look at real prices and post the result late. The models are unlikely to be posted by anyone, so there is no way to believe or verify.
Much easier to make your own signal, and show results on it.