Discussion of article "Econometric approach to finding market patterns: Autocorrelation, Heat Maps and Scatter Plots" - page 3

[Imanol Salazar Garcia]

Thank you very much for passing on such valuable information.

Reading these articles makes you want to experiment.

[Miguel Angel Vico Alba]

Imanol Salazar Garcia:

Thank you for your feedback.

I advise you to read it from the original article if you want to be read. Don't worry about the language, the translation tool will do its job.

https://www.mql5.com/ru/forum/330111

Thank you.

[Alexander Kraus]

Thanks for your work , I myself have tried to study and test patterns on chats on robots but was not able to implement such on charts. today I took up your articles and was very happy that my efforts in studying and implementing deep learning can progress further

[Nataly Tashik]

Good afternoon, please explain about the construction of the heat map:

We take the increments from the hour of one day to the hour of another day in this way:

ratesH = rates.drop(rates.index[~rates.index.hour.isin([i])]).diff(lag).dropna()

If the function is then called with a lag value of 25, we get in rateH the difference not of neighbouring days, but 25 days between days, because before that we threw out all the hours not coinciding with i, and got a series of the form 2020-01-01 00:00:00|90, 2020-01-02 00:00:00:00|87, 2020-01-03 00:00:00:00|88, and then if we apply lag 25 to it and count the diff, and then throw out na - we get a series 2020-01-25 00:00:00|1, 2020-01-26 00:00:00|3, 2020-01-27 00:00:00|-2 - and so for each hour. In the end what kind of heatmap have we constructed?

I may be confused somewhere, please help me to unravel.

[Cassio Bittencourt]

Oi

Vi que você entende bem de cointegração

Gostaria de perguntar se você sabe se esse indicador existe no mql5, ou quem criou a base para ele.

att

Cássio