Machine learning in trading: theory, models, practice and algo-trading - page 380
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
and the forecast will be "Hooray!
I think they write that "YES" is a good predictor as long as the volatility is not high.
How do I do this?
Can I see the code?
Follow the link on the previous page.
I do not understand anything.
There is a "Strategy" chart and a "Buy and Hold" chart. And where is the chartof EURUSD?
I do not understand anything.
There is a "Strategy" chart and a "Buy and Hold" chart. Where is theEURUSD chart?
Pardon me, went to the next post and found the following chart.
Very interesting stuff.
What's missing is tests that justify the applicability of the functions in question.
Because it has already made sense, and MO is just an MO. I bought a college textbook, it has a Garch and I'm reading it.)
MO is always a ready-made meaningful model. Sometimes it is so meaningful, that it is hard to understand how it works. Here's an article about gradient bouncing for examplehttps://habrahabr.ru/company/ods/blog/327250/ There is an article and there are descriptions and formulas, but I've never been able to realize my wish to transfer it into mql, it's too complicated.
The problem is a little different, not in the sense, but in the narrow specialization.
Arima and Garch work directly with prices without indicators and TA. There is an algorithm that transforms a price series into a stationary vector. They even have some mechanism that corrects forecasts based on previous errors (MA component). But they are useless for other (not price) data, these models can't classify pictures.
If we pass a time series of prices to the neural network for training, it will not search for autocorrelations, seasonal and trend components of the price, as the neural network is not able to do. It will simply remember what it was given, and for new data during a test or real trading, it will "remember" similar price vectors from the past, and trade as it was before, and this in forex means a loss.
Neuronka needs help in predicting the price - first of all it should find the indicators that, like Arima, will be able to determine autocorrelation, trend, seasonality, and the values of these indicators to send to the neuronka. Then it will have at least a small chance to be comparable to arima and garch.
Another important thing is that arima makes predictions based on time. This model remembers exactly in what order the prices arrived and uses a sort of sliding window in its forecast, taking several latest prices and making forecast by them. Unlike neuronics, which works with the whole training table at once without any idea in what order the prices came.
MO is always a ready-made meaningful model. Sometimes it is so meaningful, that you don't even understand how it works. Here is an article about gradient bouncing for examplehttps://habrahabr.ru/company/ods/blog/327250/ There is an article, there are descriptions and formulas, but my desire to transfer this into mql I haven't been able to realize, it's too complicated.
It's a little different, not in the sense, but in the narrow specialization.
Arima and Garch work directly with prices without indicators and TA. They have a built-in algorithm for turning a price series into a stationary vector, and there are even some subtleties such as correction of predictions depending on previous errors (MA component). But they are useless for other (not price) data, these models can't classify pictures.
If we pass a time series of prices to the neural network for training, it will not search for autocorrelations, seasonal and trend components of the price, as the neural network is not able to do. It will simply remember what it was given, and for new data during a test or real trading, it will "remember" similar price vectors from the past, and trade as it was before, and this in forex means a loss.
Neuronka needs help in predicting the price - first of all it should find the indicators that, like Arima, will be able to determine autocorrelation, trend, seasonality, and the values of these indicators to send to the neuronka. Then it will have at least a small chance to be comparable to arima and garch.
Another important thing is that arima makes predictions based on time. This model remembers exactly in what order the prices arrived and uses a sort of sliding window in its forecast, taking several latest prices and making forecast by them. In contrast to neuronics, which works with the whole training table at once with no idea in what order the prices came.
MO is not a finished meaningful model, it's a process :) and garch is a model. How can it even be compared :)
I'm well aware of what's going on, to some semblance of garch I went before I even knew it existed. And what neuronics does. I've been building my thoughtful model for a long time :) how long, about 2 weeks. All these articles are mostly rubbish, by the way... for general education, it's interesting to read.
And by the way, the complexity of the system in the forex market does not characterize its effectiveness in any way...How can it even be compared :)
MO always includes some model, and almost always takes some working time-proven model. For example neuronics or gradient boosting, so much effort and time has been invested in their creation (decades of algorithm evolution) that they can be even more meaningful than arima.
MO always includes some model, and almost always takes some working time-proven model. For example neuronics or gradient boosting, so much effort and time has been invested in their creation (decades of algorithm evolution) that they can be even more meaningful than arima.
What model is there - regression or classification :) That's all neuronics can do.
MO is always a ready-made meaningful model. Sometimes it is so meaningful, that it is hard to understand how it works. Here is an article about gradient bouncing for examplehttps://habrahabr.ru/company/ods/blog/327250/ There is an article, there are descriptions and formulas, but my desire to transfer this to mql I haven't been able to realize, it's too complicated.
It's a little different, not in the sense, but in the narrow specialization.
Arima and Garch work directly with prices without indicators and TA. They have a built-in algorithm for turning a price series into a stationary vector, and there are even subtleties such as correction of predictions based on previous errors (MA component). But they are useless for other (not price) data, these models can't classify pictures.
If we pass a time series of prices to the neural network for training, it will not search for autocorrelations, seasonal and trend components of the price, as the neural network is not able to do. It will simply remember what it was given, and for new data during a test or real trading, it will "remember" similar price vectors from the past, and trade as it was before, and this in forex means a loss.
Neuronka needs help in predicting the price - first of all it should find the indicators that, like Arima, will be able to determine autocorrelation, trend, seasonality, and the values of these indicators to send to the neuronka. Then it will have at least a small chance to be comparable to arima and garch.
Another important thing is that arima makes predictions based on time. This model remembers exactly in what order the prices arrived and uses a sort of sliding window in its forecast, taking several latest prices and making forecast by them. In contrast to neuronics, which works with the whole training table at once with no idea in what order the prices came.
+++