{"id":492,"date":"2017-09-24T05:53:54","date_gmt":"2017-09-24T05:53:54","guid":{"rendered":"http:\/\/jsr.isrt.ac.bd\/?post_type=article&p=492"},"modified":"2017-09-24T05:54:08","modified_gmt":"2017-09-24T05:54:08","slug":"note-construction-generalized-tukey-type-transformations","status":"publish","type":"article","link":"http:\/\/jsr.isrt.ac.bd\/article\/note-construction-generalized-tukey-type-transformations\/","title":{"rendered":"A note on the construction of generalized Tukey-type transformations"},"content":{"rendered":"
One possibility to construct heavy tail distributions is to directly manipulate a
\nstandard Gaussian random variable by means of transformations which satisfy
\ncertain conditions. This approach dates back to Tukey (1960) who introduces the
\npopular -transformation. Alternatively, the -transformation of MacGillivray
\n& Cannon (1997) or the -transformation of Fischer & Klein (2004) may be used.
\nRecently, Klein & Fischer (2006) proposed a very general power kurtosis transformation
\nwhich includes the above-mentioned transformations as special cases.
\nUnfortunately, their transformation requires an in\fnite number of unknown parameters
\nto be estimated. In contrast, we introduce a very simple method to
\nconstruct exible kurtosis transformations. In particular, manageable “superstructures”
\nare suggested in order to statistically discriminate between -, – and
\n-distributions (associated to -, – and -transformations).<\/p>\n