Machine learning in trading: theory, models, practice and algo-trading - page 2700
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
Perhaps it is relevant for multiple points, for example to find similar patterns, but in my case there is essentially one point at the first stage. The point is converted/normalised into different relative measurement systems - time scale and price, plus a third space - any discrete predictor continuously describing the market. You get 3 dimensions in the initial representation. Each has its own quantum table.
How can you normalise one point if it's one point????
Normalisation with respect to spatial predictors, not points on the same plane. Or is that not called normalisation? I'm not good with terms.
A pattern may be present, but its description by the rule will be insufficient, for example, it fits into the range from 50 to 100 bars, and the rule sets the average value of the limits.
I don't get it
Well, I'm not ready to reveal the whole kitchen in detail yet. I'm looking for people who want to move with me.
it's a known fuck up... and it's always knocking things down, no matter what we trade:-) in the two main centres - USA and England the clocks are set on different days. Up to more than 1 week apart. The intervals between the most important events change and two or three weeks in six months can be thrown out of the analysis. And ours are still making a mess of things, "we change the clocks, we don't change them".
I don't know a universal or even more or less successful solution to this problem. Either just ignore these "critical days" or teach winter/summer time separately. The latter seems more reasonable, but we're already critically short of data as it is
We need to think about this. I've tried shifting the time - the result was clearly positive - time-related patterns were preserved. The trouble is that when there are many such predictor spaces, the model cannot always use them all, and it is not easy to teach it to think in this way.
Well, I'm not ready to reveal the whole kitchen in detail yet. I'm looking for those who want to move with me.
It's a revelation, I didn't know about it.
Or maybe we can make a joint project a la MQL5, only in Python, designed for creating bots.
Django and quotes from Binance can be used as a basis, indicators are available. We will collect a personal account of the user, which will be a prototype of the terminal, there is a bible for rendering in the browser.
Now all the dice have been invented, it remains to collect them in one place. And the motivation is good - MQL5 does not work with crypto, and this project will be purely for crypto.
Well, that's just... fantasy. Don't throw your fists at me.