Machine learning in trading: theory, models, practice and algo-trading - page 1828
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
synthesize rules that take into account strictly some specific events in a specific sequence, as well as take into account the price at which these events occurred.
Here is a certain pattern
Here is this pattern in prices
here is the same pattern
The problem here is similarity without correlation. This is where the metric of economics began) Rows are similar, but there is no correlation between them. I like the idea, it's more like creating a library of states, but purely incremental lags are not enough, we need more predictors, if at all. Nobody forbids to find repeating combinations, if these combinations are distributed on large sections more or less evenly, there is a chance.
the problem here is similarity without correlation. This is where economics metrics began) Rows are similar, but there is no correlation between them. I like the idea, it is more like creating a library of states, but purely incremental lags are not enough, we need more predictors, if they exist at all. Nobody forbids to find recurring combinations, if these combinations are distributed on large intervals more or less evenly, there is a chance.
What lags... you will kill the price !
A state is a rule, it could be anything - a candlestick combo, a level breakout, the level itself, a volatility kickout, whatever
The states in the right order create a pattern of any size.
No BP forecasting algorithm is able to look for such patterns
What lags... you will kill the price!
A state is a rule, it can be anything - a candlestick combo, a level breakout, the level itself, a volatility kickout, whatever
the states in the right order create a pattern
Anything )))) a candlestick combination, a breakout, a return to the mean, a trend change or an extremum, an ejection, an ejection/volatility change. Nothing else comes to mind that can be described mathematically.
To add a major and minor TF or several more ones may be necessary. The idea is clear at the top, but the stationary sequences have not been found yet. Although usually look for the maximum on the target and junior.
No BP prediction algorithm knows how to look for such patterns
That's how you set the recognition task. BP prediction and image recognition from incomplete data are different tasks.
This is how you set the task of recognition. BP prediction and image recognition from incomplete data are different tasks.
The market is not BP
the market is not BP
Cool) what then. Values at a point in time is a series.
I do not know ...
You can represent anything as a time series, there is always time.
But if no BP prediction algorithm can predict the price in any way, isn't the answer obvious?
I don't know...
You can imagine anything as a time series, there is always time.
But if no BP prediction algorithm can predict the price in any way, isn't the answer obvious?
Time can be replaced by a number.
What will it do? Nothing.
At least the direction or the range was already good. It's possible to make a forecast within the range of stationarity. The problem is in the determination of the stationarity boundaries only by history, the problem is without a 100% forecast, but it's not zero, either.
It's all working with BP again, it won't work