6 Comments

Always an informative read!

Expand full comment

Great as always!

Expand full comment

Very interesting and applicable concepts in every blog. Thank for share!

Expand full comment

Very well explained. Thank you

Expand full comment

Well I finally understand what a QQ plot is showing, awesome! Quick question - why would you choose say a 25-7th percentile reference line as opposed to a regression line? What influences that decision and what impact does that decision have on evaluation? Also - what is the reference line generated by default in say R or python?

Expand full comment

Considering the regression fit as the reference line can be influenced by outliers. That is why we prefer the line connecting 25-75th percentile. I am not sure about R but in Python, we must specify the parameter when creating a QQ plot, i.e., there is no default value: https://www.statsmodels.org/stable/generated/statsmodels.graphics.gofplots.qqplot.html

Expand full comment