Machine learning in trading: theory, models, practice and algo-trading - page 2275
![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
How about DTW?
An article on the Hubra about the use of DTW in speech recognition.
No, DTW is for local areas, plus it's slow...
And they'd have to be in a sliding window to walk all over the graph + with all the dimensions of the sliding window...
that's not an option...
I'm sure there is something brighter...
Grand sliding normalization + correlation
abra + cadabra
Thanks all clear)
It's on amplitude, and he seems to want it on time.
He wants all of it.
The algorithm should understand that it's all the same
There's a difference in time and amplitude and frequency....
The same ONE is the proportions between the dots in the pattern
This is in amplitude, and he seems to want to do it in time.
great resampling by time, break into wider bars (bins), or vice versa interpolate
interpolation seems preferable, because there will be no missing infoWell, here' s an example of how scientists look for patterns. But we are much smarter and our bikes are much better.
I repeat "for those in the tank" TDW for locally similar patterns
You're not going to swivel the mouse in the area of 20-100 meters around yourself, all the patterns that are in the article will be within 0.5 meters at most... (they are locally similar (in time) )
I, on the other hand, need complete invariance in size
I repeat "for those in the tank" TDW for locally similar patterns
You won't be swinging your mouse in the area of 20-100 meters around you, all the patterns in the article will be within 0.5 meters at most... (they are locally similar (in time) )
I, on the other hand, need complete invariance in size.
That's okay, in a dozen years you'll discover logarithm.
I repeat "for those in the tank" TDW for locally similar patterns
You won't be swinging your mouse in the area of 20-100 meters around you, all the patterns in the article will be within 0.5 meters at most... (they are locally similar (in time) )
I, on the other hand, need complete invariance in size
The logic of the range size, from which to look for scales, or rather in what scale to look for a pattern. Otherwise, by a complete enumeration of windows from the width of the maximum pattern to the minimum, and so sliding over a number of expensive. but surely all will be found. the width of the maximum pattern only need to understand.
That's okay, in about ten years you'll discover logarithmetic.
How will logarithmetic help to recognize the patterns I drew above?
The logic of the spread value on the range, from it already look for the scale, or rather in what scale to look for a pattern. Otherwise, by the full search of windows from the width of the maximum pattern to the minimum, and so sliding on a number of expensive. but surely all will be found. the width of the maximum pattern only need to understand.
You realize it'll take half a day to find one pattern, a complete search...
I think there's a more elegant way