Normal or not normal?

Most often tests for normality are not robust (a presence of outliers affect the results and data significantly deviate from normality). On the other hand, robust methods of normality testing [1] can deal with it. Another article, very worthy of reading for everyone.

 

Share on FacebookShare on LinkedInShare on Google+Tweet about this on Twitter

References

  1. M. Stehlík, L. Střelec, and M. Thulin, "On robust testing for normality in chemometrics", Chemometrics and Intelligent Laboratory Systems, vol. 130, pp. 98-108, 2014. http://dx.doi.org/10.1016/j.chemolab.2013.10.010

Comments

comments

Comments are closed