Machine learning in trading: theory, models, practice and algo-trading - page 2712

New comment
 
Evgeny Dyuka #:
Is this the best and only forum on "Machine Learning in Trading" in runet or are there others?
This is far from the best, it is the only one
 
Aleksey Vyazmikin #:

Alexey thinks right, about events and correct sequence of events, but without theory on algorithms I think it all will remain at the level of talk.

=============

Here is an example of a pattern

1) the left edge of the chart up to the red vertical line is a pattern with a level.

2) the first blue arrow is the correct reaction to the pattern

3) again the correct reaction to the pattern is the entry point.

As you can see the pattern is the same, it works "my honour", can it be found by looking at the market as a time series ??? is a rhetorical question.

I want to see someone who can find such patterns with neuron or catbustom by moving onknock and seeing the last n candles.

 

Calling examples for learning as events, a person substitutes notions as if it is something that really happened. But these are his invented examples, which cannot be called events.

Then the person starts to believe that these are events, not training examples taken from the ceiling, substituting reality for fakes with his definitions

Then he starts living in this fake reality, citing some stupid diagrams to justify it, which have nothing to do with anything.

Then he turns out everyone is a fool, calls to expand consciousness, etc. etc.

And then the financial time line turns into something sacred and eventful.

 
The plane took off and landed, recorded the event. We entered an altered consciousness and saw that the plane crossed the RSI calculated by its trajectory. That's also an event for us. Now we need to calculate the difference between an adequate person and a mantis.
 
Let us imagine that you are a mantis and cannot distinguish one fictitious event from another. For you, cause and effect are one event and, therefore, occur simultaneously without any temporal and informational connections between them, since there is no time series and the events themselves are fictitious. You collect such fictitious events, wrap them in leaves, admire your work, tell your friends about them. Your events happen outside of time and space in your head. And you seem to be doing pretty well, but all of this bears little resemblance to machine learning.
 
Do you have on git hub, project, at least see what the heated arguments are about here?
 

What does it all mean? Do Ilearn tensorflow for nothing?


  1. MQL5: Added Activation (activation function) and Derivative (derivative of activation function) matrix and vector methods with parameters:
    .
    AF_ELU Exponential Linear Unit
    AF_EXP Exponential
    AF_GELU Gaussian Error Linear Unit
    AF_HARD_SIGMOID Hard Sigmoid
    AF_LINEAR Linear
    AF_LRELU Leaky REctified Linear Unit
    AF_RELU REctified Linear Unit
    AF_SELU Scaled Exponential Linear Unit
    AF_SIGMOID Sigmoid
    AF_SOFTMAX Softmax
    AF_SOFTPLUS Softplus
    AF_SOFTSIGN Softsign
    AF_SWISH Swish
    AF_TANH Hyperbolic Tangent
    AF_TRELU Thresholded REctified Linear Unit
    The activation function in a neural network determines how the weighted sum of the input signal is converted into the output signal of a node or nodes at the network level. The choice of activation function has a great impact on the capability and performance of the neural network. Different activation functions can be used in different parts of the model. MQL5 implements not only all known activation functions, but also derivatives of the activation function. Derivative functions are needed to quickly calculate the correction based on the received error during the training of the neural network.
  2. MQL5: Added Loss matrix and vector method (Loss function) with the following parameters:

    .
    LOSS_MSE Mean Squared Error
    LOSS_MAE Mean Absolute Error
    LOSS_CCE Categorical Crossentropy
    LOSS_BCE Binary Crossentropy
    LOSS_MAPE Mean Absolute Percentage Error
    LOSS_MSLE Mean Squared Logarithmic Error
    LOSS_KLD Kullback-Leibler Divergence
    LOSS_COSINE Cosine similarity/proximity
    LOSS_POISSON Poisson
    LOSS_HINGE Hinge
    LOSS_SQ_HINGE Squared Hinge
    LOSS_CAT_HINGE Categorical Hinge
    LOSS_LOG_COSH Logarithm of the Hyperbolic Cosine
    LOSS_HUBER Huber
 
mytarmailS #:

Alexey thinks right, about events and the correct sequence of events, but without the theory of algorithms I think it all will remain at the level of talk.

=============

That's right, I may need algorithms so that people understand how they work - without that you can't code them.

Unfortunately, I can't even read complicated formulas - I don't understand them.

Regarding algorithms, I remember a case from my youth, I was in college, I think the first year, it was computer science, we were learning Basic. Classes were held in a computer class, my friend and I entertained ourselves during the classes by snooping around the local network, sniffing computers and other young entertainment.... we were sitting far away from the blackboard and the teacher, so I didn't see anything there, so sometimes I could hear what they were talking about.... It was time for a test or a lab - I don't remember, I had to write a programme on the computer that would calculate something and display it on the screen. It is clear that the teacher did not like me, so when I finished the work early, and again entertained, she sat down at my computer ran through the code and maliciously told the audience that I am a talentless and my code sucks and it will not work, and what a raztakaya me .... So the people gathered, and she ran the code, and it worked! And for the rest of the class she just sat there denying reality. That was the last session of our group with her - she refused to work with us because of me," the supervisor told me.

Then, after a couple of months, there was a man who turned out to be a great man, and allowed even to score on other pairs, secretly sitting in his office and hanging around on the Internet, taking home on floppy discs dug up software from the web.

Eh, it was a long time ago...

 
Maxim Dmitrievsky #:

Calling examples for learning as events, a person substitutes notions as if it is something that really happened. But these are his invented examples, which cannot be called events.

Then the person starts to believe that these are events, not training examples taken from the ceiling, substituting reality for fakes with his definitions

Then he starts living in this fake reality, citing some stupid diagrams to justify it, which have nothing to do with anything.

Then everyone turns out to be a fool, calls to expand consciousness, etc., etc.

And then the financial time line turns into something sacred and eventful.

Maxim, vent your bile, relax - you don't have to prove your superiority and uniqueness at all!

I have not insulted anyone - try to find it.

If your vision is a matter of religion and faith, then I dare not interfere by trying to let you see more than time series.

 
Maxim Dmitrievsky #:
The plane took off and landed, recorded the event. We entered an altered consciousness and saw that the plane crossed the RSI calculated by its trajectory. That's also an event for us. Now we need to calculate the difference between an adequate man and a mantis.

That's a good example, we can look at that. I really don't understand much about aeroplanes, but this is just an illustration of the essence - it may help you to understand the essence.


1...270527062707270827092710271127122713271427152716271727182719...3525
New comment
Reason: