0
$\begingroup$

Let's say you have online Profile A. Profile A is present on 3 websites: X, Y, Z.

Profile A on site X contains the following fields:

name: A
email: a@a.com
sex:m
age:35
location:europe

Profile A on site Y contains the following fields:

name: B
email: a@a.com
sex:f
age:28
location:us

Profile A on site Z contains the following fields:

name: A
email: b@b.com
sex:t
age:35
location:asia

This is just an example it could be any other kind of data.

My question is about whether a threshold exists of recordset randomness. How would that be detected?

Where models will not derive anything significant out it. Kinda of like poop in -> poop out kind of scenario.

What is this even called in the field of Data Science? So I can read more about it.

Does this kind of data just get ignored and become irrelevant?

Does it result in heavily unrealistic estimations?

Does it solely depend on how the model is structured whether meaning is derived from this example?

Thanks

$\endgroup$
2
  • $\begingroup$ I am not sure I understand the question. Data may be meaningless or junk. On the other hand ML models if forced can find patterns even in noise (irrelevant patterns for sure) $\endgroup$
    – Nikos M.
    Nov 17 '21 at 21:21
  • $\begingroup$ @NikosM. So even if the data has a certain tendency depending on the the types of occurrences it contains. There isn't anyway to verify the models assumptions. Because the model is using the data to paint a picture vs the other way around. So it's almost like any story/narrative can be crafted from any dataset given enough data... Regardless if the data is super random or not. $\endgroup$
    – Dan
    Nov 17 '21 at 22:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.