Machine learning in trading: theory, models, practice and algo-trading - page 1708

[Реter Konow]

Valeriy Yastremskiy:

Probably not so after all. And Hegel was sometimes wrong. And he was too sure of his erroneous assertions. It must have been an erroneous certainty.

Newton's "famous decomposition of forces" testifies to Hegel's "complete lack of a sense of nature".

Also, Hegel wasn't wrong about many things))))

Nowadays only lazy does not earn on a virus, and since times are hard, this is a moral justification for such earning)

Yes, I agree.

[Aleksey Nikolayev]

Reg Konow:
The astronomical conclusions are wrong. But, to "juggle" authority and the importance of his personality in general, using such a primitive reason can only a person with great conceit. In short, a foolish man.

This is your speculation. I only advised you not to become as much of a "biologist" as Hegel was an "astronomer. I even found you a link to an article about the study. But reading articles is not your method, is it?

[Valeriy Yastremskiy]

Aleksey Nikolayev:

This is your speculation. I only advised you not to become as much of a "biologist" as Hegel was an "astronomer. I even found you a link to an article about the study. But reading articles is not your method, is it?

To be honest I did not understand at once that it is a reference to the article. And this article is purely on the subject of MO, not in trading, but the methods are considered in detail.

[Реter Konow]

Igor Makanu:

it is not speculation, but a limited vocabulary, if i am not mistaken, it is recommended to children for this reason and to read as much literature as possible, but i do not know how effective it is for an adult )))

juggle with authority, rather ambiguously phrased, I will not gulp, but imho - the authority is something permanent, immutable, and how this here is possible to juggle?

In the stock from the forum verbal battles: "you do not repaint.

Before you talk about the limitations of other people's vocabulary, you should at least check the punctuation and spelling of your post.)) I can, too. I have been living in a non-Russian-speaking environment for 25 years, which, of course, sometimes affects the vocabulary. But, I try, unlike some.

And the expression "juggling with authority" very accurately describes how the name of a great man is used by an unknown Nikolaev.

[Реter Konow]

Aleksey Nikolayev:

This is your speculation. I only advised you not to become as much of a "biologist" as Hegel was an "astronomer. I even found you a link to an article about the study. But reading articles is not your method, is it?

I took a cursory look at the original article. Only an expert can thoroughly understand it. I just expressed my doubt that the article on "Zen" is objective and doesn't distort the reality by wild exaggeration.

[Igor Makanu]

On the topic of the topic

If anyone has a ready example of RNN-network based on LSTM for the price?

can be on Python, you need the order of networking and training to look, on the Internet examples are tons, but on the subject of the material to be studied (CD) I want to look at

Thanks

[Реter Konow]

According to an article on Zen, the AI selects successful chemical compounds against bacteria only by statistical analysis of data and cause-effect relationships. The algorithm does not use empiricism, i.e., it does not test by experience and yet, its result is highly accurate. This is odd.

If this method works, then successful antibacterial drugs can be quickly and effectively withdrawn, without much experimentation. But, the current viral environment refutes such a conclusion.

[Aleksey Nikolayev]

Retag Konow:
Judging from the Zen article, the AI selects successful chemical compounds against bacteria, just by statistical analysis of data and cause-effect relationships. The algorithm doesn't use empiricism, i.e., it doesn't test by experience and yet, its result is highly accurate. This is odd.

If this method works, then successful antibacterial drugs can be quickly and effectively withdrawn, without much experimentation. But, the current viral situation refutes such a conclusion.

You can't do without experiments. The main idea, as far as I understand it, is simply to significantly reduce the list of substances allowed for experiments. Here is a link to a more meaningful story in Russian about this research, with an emphasis on biology and without the details of the MO.

[Igor Makanu]

Igor Makanu:

On the topic of the topic

If anyone has a ready example of RNN-network based on LSTM for the price?

can be on Python, you need the order of networking and training to look, on the Internet examples are tons, but on the subject of the material to be studied (CD) I want to look at

Thank you

Question removed, I remembered and found where I was reading, everything is in the articlehttps://www.mql5.com/ru/articles/3473

[Forester]

Aleksey Vyazmikin:

Alexei, you do leaf analysis, apparently you can answer... Or someone who knows.

Here's a description of the splits of a tree as deep as 2 of kaboos

  {
    "left": {
      "value": -0.5163727959697733,
      "weight": 385
    },
    "right": {
      "left": {
        "value": -0.40476190476190477,
        "weight": 156
      },
      "right": {
        "value": 0.18232044198895028,
        "weight": 350
      },
      "split": {
        "border": 4.999999046325684,
        "ctr_target_border_idx": 0,
        "split_index": 4,
        "split_type": "OnlineCtr"
      }
    },
    "split": {
      "border": 12.762499809265137,
      "float_feature_index": 1,
      "split_index": 1,
      "split_type": "FloatFeature"
    }
  },

What does "value" mean? Is it a leaf answer? What do negative numbers mean?

If so, what is value for multiclass classification? Below are splits of one of the trees trained on 3 classes.
At each leaf we see an array of 3 values of value. What is the answer? The largest value? Then why store redundant two values? What do the negative values mean?

[
  {
    "left": {
      "left": {
        "value": [
          0.8474576271186454,
          -0.4237288135593223,
          -0.4237288135593223
        ],
        "weight": 50
      },
      "right": {
        "value": [
          0,
          0,
          0
        ],
        "weight": 0
      },
      "split": {
        "border": 6.850000381469727,
        "float_feature_index": 0,
        "split_index": 0,
        "split_type": "FloatFeature"
      }
    },
    "right": {
      "left": {
        "value": [
          -0.4262295081967212,
          0.7540983606557395,
          -0.3278688524590165
        ],
        "weight": 52
      },
      "right": {
        "value": [
          -0.42105263157894807,
          -0.3684210526315795,
          0.7894736842105271
        ],
        "weight": 48
      },
      "split": {
        "border": 1.6500000953674316,
        "float_feature_index": 3,
        "split_index": 4,
        "split_type": "FloatFeature"
      }
    },
    "split": {
      "border": 0.800000011920929,
      "float_feature_index": 3,
      "split_index": 3,
      "split_type": "FloatFeature"
    }
  },

Interestingly, the sum of the three values is 0.