Machine learning in trading: theory, models, practice and algo-trading - page 398
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
Looked at the file, there are 8 coefficients input is23 i.e. there is a neuron to work with 3 inputs. I guess, that twenty-four hours your program counts, which 3 inputs out of 100 to submit to this neuron. I thought there neuron is expanded to at least 10 inputs...
Yes, there is a rigid selection, only a few are left :)
And I sent the link to his website, there is a description of the model. I don't even know how to describe it, it says Nuclear + Vector Machine. It's more complicated there than in MT5 + training with the opponent instead of selecting weights in the optimizer, but in the output we have the same weights for each of the predictors.
As far as I know the method of reference vectors is used, when in multidimensional space the hyperplane is built which is maximally removed from all points of training set, thereby dividing the space into two categories "yes" and "no". Spitting kernel transformations. So what takes the optimizer VERY long is not a neuron to train, it's an AI system, which in addition to training does a bunch of all sorts of transformations and brute forcing with data. That's why it takes so long. Which is a bit frustrating. But then again, how can you get a model adequate to the market in 5 seconds???? I don't know :-(
As far as I know, the method of reference vectors is used, when in multidimensional space one constructs such a hyperplane which is as far as possible from all points of the training set, thus dividing the space into two categories "yes" and "no". Spitting kernel transformations. So what takes the optimizer VERY long is not a neuron to train, it's an AI system, which in addition to training does a bunch of all sorts of conversions and data manipulations. That's why it takes so long. Which is a bit frustrating. But then again, how can you get a model adequate to the market in 5 seconds???? I don't know :-(
Well at least, as it was already written above - don't shove a bunch of predictors into it, it will eliminate them all anyway :)
Looked at the file, there are 8 coefficients input is23 i.e. there is a neuron to work with 3 inputs. I guess, that twenty-four hours your program counts, which 3 inputs out of 100 to submit to this neuron. I thought that neuron is expanded to at least 10 inputs...
getBinaryClassificator1(x0, x1, x2, x3, x4, x5, x6, x7);
Not something you misunderstood, there are 2 grids, each is fed with 8 inputs, then the result of both grids is interpreted as described before. If both say yes, then yes, if not, then no, if in different ways, then I don't know... Like this.....
That's a tough one... Is it possible to predict something from 3 inputs?
I think you're wasting your time in the wrong place. If the RNN were to expand to 10 or 100 inputs, then you could spend 24 hours on his training. And now you are wasting your time on sifting out entries. And not on training.
Not something you misunderstood, there are 2 grids for each 8 inputs, then the result of both grids are interpreted as described before. If both say yes, then yes, if not, then no, if mixed, then I don't know... Like this.....
8 are the coefficients of connections between inputs and outputs, and there are 3 inputs (8 =23)
More precisely, the coefficients for the rules:
r0 = (1 - A) * (1 - B) * (1 - C) * p0
And there are only 3 inputs - A,B and Cr1 = (1 - A) * (1 - B) * C * p1
r2 = (1 - A) * B * (1 - C) * p2
r3 = (1 - A) * B * C * p3
r4 = A * (1 - B) * (1 - C) * p4
r5 = A * (1 - B) * C * p5
r6 = A * B * (1 - C) * p6
r7 = A * B * C * p7
That's a tough one... Is it possible to predict something from 3 inputs?
I think you're wasting your time in the wrong place. If the RNN were to expand to 10 or 100 inputs, then you could spend 24 hours on his training. And now you are wasting your time on sifting out entries. And not on training.
because the latest version there uses a committee of 2 ns, then the results are compared
because the latest version there uses a committee of 2 ns, then the results are compared
I would expand the neuron to 10 inputs.
But I need to add rules up to 1024:
r0 = (1 - A) * (1 - B) * (1 - C) * p0
r1 = (1 - A) * (1 - B) * C * p1
r2 = (1 - A) * B * (1 - C) * p2
r3 = (1 - A) * B * C * p3
r4 = A * (1 - B) * (1 - C) * p4
r5 = A * (1 - B) * C * p5
r6 = A * B * (1 - C) * p6
r7 = A * B * C * p7
.....
r1023 =
Well at least, as written above - you shouldn't shove a bunch of predictors into it, it will eliminate them all anyway :)
Here I'm willing to bet. This is from practice. What I'm observing during using optimizer.
If there are less columns than rows, the number of inputs becomes less than the conditional size of the network polynomial.
If the number of columns and rows are equal, then the number of inputs and the size of the polynomial are about the same.
If there are more columns than rows, the size of polynomial is less than the number of inputs.
In other words, the number of columns and rows must be about the same. Then the polynomial is obtained with sufficient number of inputs, and its size is about the same as the number of inputs.
In other and other words. The larger the file, the more multiparametric the model becomes. The number of inputs and the size of the polynomial increase, which indicates that the model becomes smarter. So there you go.
Also, when running the same file repeatedly, the set of input parameters is always different. This is due to random partitioning of the sample during division. But if we run the same file and get totally different models, they will still work approximately the same, at least they should. This is the answer to the question of which fish are in the given file. If data differ from run to run, it says that the data has nothing to do with the output. Like this....
8 are the coefficients of connections between inputs and outputs, and there are 3 inputs (8 =23)
More precisely, the coefficients for the rules:
r0 = (1 - A) * (1 - B) * (1 - C) * p0
There are only three inputs - A, B, and C.r1 = (1 - A) * (1 - B) * C * p1
r2 = (1 - A) * B * (1 - C) * p2
r3 = (1 - A) * B * C * p3
r4 = A * (1 - B) * (1 - C) * p4
r5 = A * (1 - B) * C * p5
r6 = A * B * (1 - C) * p6
r7 = A * B * C * p7
What coupling coefficients are you talking about? Here is a list of used inputs, and what you see there before each polynomial is normalization, and there are 8 inputs described, look below at the very bottom of the file.
//Variable x0: Del
//Variable x1: Del1
//Variable x2: VDel
//Variable x3: VDel1
//Variable x4: VDel6
//Variable x5: ST5
//Variable x6: VAD11
//Variable x7: VVolum4
Where are the three inputs????? When there are eight of them.
The example you cited is from another work by Reshetov and has nothing to do with the optimizer.