Machine learning in trading: theory, models, practice and algo-trading - page 2192
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
i.e., there is practically nothing but descriptive "water," which has nothing to do with science.
yes, but even that is closer to trading than tsos
Come on, stop arguing... He who understands understands.
Perhaps understanding will come to someone else.
Yes, but even that's closer to trading than tsos.
Well, it's already upside down.
Perhaps understanding will come to someone else.
It just needs to be studied at least cursorily to understand....
to understand that any manipulation with the function is already DSP, any!!!
filtering, prediction, processing, modulation, pattern finding, analysis...
the same neural networks, they are digital filters, and digital filters are a section of DSP
it's just that everyone here has a notion that DSP is kind of radio signals and bleep
It just needs to be studied at least cursorily to understand....
understand that any manipulation with the function is already DSP, any!!!
filtering, prediction, processing, modulation, pattern finding, analysis...
the same neural networks, they are digital filters, and digital filters are a section of DSP
It's just that everyone here has this notion that DSP is like radio signals and bleep.
Yes. Just because a price (quote) is not a continuous function, it is discrete in time and quantized in level - that's why it has to be processed according to DSP rules.
Yes. Just because a price (quote) is not a continuous function - it is discrete in time and quantized in level - that is why it must be processed according to DSP rules.
I see against the trend so far. The deals for today will not be until tomorrow
no longer a week and no longer a competition....
subscriber e.
Yes. Just because the price (quotes) is not a continuous function, it is discrete in time and quantized in level, so it must be processed according to DSP rules.
Temperature is also measured discretely. Lan is good. DSP and ECM are applied sciences, not fundamental. I like the filters did not fall for, although microwave with a magnetron and klystron and comb filters seemed to count)))) Well did not fall in love at all))) But when the formula really saw here with the pros and cons of the filter, then multidirectional comb remembered)
So here's the deal
I want to predict channel beams
To build a channel, you need three parameters
1) height from tt (current point)
2) bottom from tt
3) channel slope
There are three values in total...
I need to turn these three values into one in order to train a regression model...
After that, I need to convert the predicted value into three values (channel parameters) again.
I thought about doing it with PCA, reducing dimensionality to one component and then doing inverse transformation from forecast...
But it didn't work to describe the three parameters with one component, too much information loss, one component is not enough
Karoch, we must solve the problem, who can train networks with three outputs (parameters) in regression?
Or have other suggestions?
PS. three models to train is not an option