Is there a neural network to find the correlation coefficient?

Question

Neural networks

Good day to all.
In the process of work, it became necessary to find the correlation dependence of the data series.
Example: I have a database with several tables, each of which has only 2 columns - these are our rows. The data in each table is synthesized using different algorithms unknown to me. Is there a neural network that can calculate the correlation coefficient without thinking about the nature of the origin of the data?
Forgive me for the broken presentation of the question - experience in the mat. statistics are not enough (literally, a semester at the university), but the problem needs to be solved. Perhaps I am too bothered and there are some methods that allow us to perform this task universally and with sufficiently high accuracy without the use of neural networks. If you know about those - advise the literature or at least a request to Google, please.

Answer 1

Perhaps I am too bothered and there are some methods that allow us to perform this task universally and with sufficiently high accuracy without the use of neural networks.

Exactly. The linear correlation coefficient is calculated as the covariance divided by the product of the standard deviations:
For example, numpy has a corresponding ready-made function .

Answer 2

I will add to the colleague's answer - this coefficient is applicable only if it is known that the original data have a normal distribution and if they (the data) are measured in interval or ratio scales. For data measured in other scales - order, nominal, dichatom - other analogues of this formula are used.
Literature - if for reference - Kobzar AI "Applied Mathematical Statistics".
If for study - any book on mathematical statistics, where there is a section "correlation analysis".