What exactly do you need? No, because it is so basic I don’t know any research paper on the ZeroR classifier?

]]>Also, is there a research paper on Zeror classifier? If so, please provide link.

Please get back to me soon. ]]>

yeah, thx

]]>The infection rate is the presumed true infection rate in this scenario, not an inferred one.

Yes, your scenario is basically my second example. Compared to the overall population it is biased towards a higher infection rate.

]]>No, sick people are divided into “true positives” and “false negatives”. To get the number of false negatives you just multiply by one minus the true positive rate (= false negative rate) and multiply this by the infection rate (= rate of sick people).

Does that make it clear?

]]>using the infection rate implicitly assumes the test is applied to individuals randomly sampled from the whole population?, am I right?

If, for example, the test would be applied to those that select themselves for the test (presumably because they had mild symptoms), the calculations as shown in the post would not apply anymore, right?, .., in that case you would have an overall probability of have being infected and a lower probability of false positives?

]]>`FN = (1-TPR)*IR`

Is it applying bayes theorem?, …, I didn’t follow, sorry

]]>Thank you, Mauro.

The units depend on the definition and context of the respective measure, e.g. with degree centrality it is the number of nodes (e.g. the number of followers in the context of social media).

There are generalizations for weighted graphs, see e.g.: Node centrality in weighted networks: Generalizing degree and shortest paths.

]]>Thank you, highly appreciated! You can find me here:

LinkedIn: https://de.linkedin.com/in/vonjd (please click “Follow” if we don’t know each other or haven’t interacted in any way, if you want to connect please add at least a note)

Twitter: vonjd @ephorie

Cheers,

Mauro