Machine learning in trading: theory, models, practice and algo-trading - page 2714
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 didn't expect such a difference in views and approaches))))
The first thing that came to my mind when I got acquainted with the terminal in 18, was to take news and global events, digitise them in a formalised way and play with the history. Heh heh, the weather is easier.
Of course, in essence, having only price data, it makes no difference what events or aggregate or sum of events caused this or that price behaviour. I still don't understand Alexey, why we should look for the cause of price movements if we can't link FA and prices yet. Weather also has current data and the task of forecasting, the causes of current data are not searched for, maybe something global is searched for, medals for the sake of course, but it has nothing to do with forecasting. Without formalisation of the current state of the market in the form of news, background, money in circulation, state of economy and industry, the task and terminology is not obvious. Of course, we understand that everything has reasons in the form of events, but this is a speculative superstructure over the data, moreover, without understanding the links between events and prices, the probability of errors is more than and this superstructure will not be able to improve the forecast, which is the target in different variants.
About terms. An event is still a single action. In the environment there is always a set or sum of events. Since sometimes there is, although I do not like this term, synergy, which can result in avalanche-like consequences or amplified many times, the sum of events is not correct. I prefer the sum of events. And at the price it is still a pattern.
The task of course is noble and even logical to connect price movements with the real world, but not today. The curse of dimensionality requires trillions of times more power, we just have to wait))))) And today it is still better to look for more down-to-earth and reachable things))))) And one should know the terminology of neighbours at least for communication)))))
because it's called time series classification, there are no other terms for it.
how it's done is irrelevant. You collect data, do any manipulation (which are not events) in order to make predictions
it's all data, not events.
How smoothly the discussion has slipped into terminology - apparently under the crust you are becoming aware, but for now the vagaries of the "I" are still preventing you from accepting reality.
For me, theory is a toolbox. If my tool doesn't fit in the "box", I can safely put it next to it or even in another one.
This does not prevent me from taking tools from different drawers and doing the same job with them.
You probably know that, for example, in China there were 365% in the circle, and it did not prevent them from inventing.
That's the point, you have to do data transformation, looking for the right formulas/algorithms, and these formulas/algorithms I called events, because they affect the price in the trader's mind/algorithms.
I did not ask for help. Mantis can think up any abstractions he wants, but it all happens in a parallel reality.
This is my nature - I always want to help those who suffer....
What a hard time people have with elementary concepts, can't tell the difference between events and their mental constructs. It's a mess. A pattern is not an event either, because it's in your head.
Everyone has a different world in their head because everyone has different degrees of receptor sensitivity and life experiences.
My approach takes this into account by allowing the fantasies of the generators of the trading process to be incorporated into the system.
Everyone has a different world in their head because everyone has different degrees of receptor sensitivity and life experiences.
My approach takes this into account, allowing the fantasies of the generators of the trading process to be incorporated into the system.
I still don't understand Alexei, why look for the cause of price movements if we can't link FA and prices yet.
One of the reasons is to increase the forecasting horizon, because if we can determine the cause with a high probability, we can immediately take into account the outcome. As a consequence, it will allow to position TP and SL more accurately, which will raise the expectation matrix and make a working tool out of a test grail.
You have a worthy opponent here, just at your level of development, communicate with him
Kindergarten across the street, go ask what your behaviour is called.
... No, I'll give you a chaperone, there's still a lot of cars.
Kindergarten across the street, go ask about the name of your behaviour.
... No, I'll give you a chaperone, there's a lot of cars, you never know...
Well, one fool was given a catbust, for example... she got into it, started to check something there, create events and quantise. I'd better not write about bousting at all then 😀 as they say, I gave birth to you and I'll kill you.
There are cubes across the street, it's hard to take them apart, so the contents won't scare you...
There are cubes across the street, they are difficult to take apart, so the contents will not scare you....