“Œv”—Œค‹†Š

M.C. JonesŽ‚ฦArthur PewseyŽ‚ษ‚ๆ‚้ƒZƒ~ƒi[

“๚Žž
2013”N3ŒŽ26“๚i‰ฮj 14:00`16:00
“o˜^•s—vฅŽQ‰ม–ณ—ฟ
๊Š
“Œv”—Œค‹†Š ƒZƒ~ƒi[Žบ‚Si‚RŠKD312Bj
yu‰‰1z 14:00`15:00
u‰‰Žา
M.C. Jones (The Open University, UK)
u‰‰‘่–ฺ
Extending univariate families of distributions to the bivariate case
ŠT—v
First, some families of univariate continuous distributions with just a very few shape parameters controlling, for example, skewness will be considered briefly. A popular question following talks wholly on such a topic is "Do these families extend naturally to the multivariate case?".
This question is considered in this talk, largely, for simplicity, in the bivariate case. It all depends, of course, on what one means by "naturally" and, indeed, what one desires of a multivariate extension. Taking just one possible course, with an emphasis on the univariate distributions as marginals (as opposed to conditionals) for the variables themselves (as opposed to combinations thereof), leads inexorably, through both obvious and less obvious specific extensions, to a major role for copulas. A particular line of work in progress drawing parallels between bivariate elliptical and Archimedean copulas will then be pursued.
yu‰‰2z 15:00`16:00
u‰‰Žา
Arthur Pewsey (University of Extremadura, Spain)
u‰‰‘่–ฺ
Circular statistics in R and beyond
ŠT—v
Over the last year I have had, perhaps to the annoyance of some of my research collaborators, a highly entertaining time co-authoring the book "Circular Statistics in R". In my talk I will: illustrate how R, its "circular" package, and the CircStatsinR workspace can be used to analyze circular data; provide details of some of the distributions available for modeling circular data; describe some of the potential research themes identified during the production of the book.
ฃ ‚ฑ‚ฬƒy[ƒW‚ฬƒgƒbƒv‚ึ