Welcome to KIT WITHERS' Home Page

I'm a research statistician at the Applied Mathematics Group of the Industrial Research Ltd, Lower Hutt, New Zealand. If you want an idea of some of my research interests you could look at this latex file of my publications. This includes extending cusum theory to track functionals other than the mean, reducing inference error, bias reduction, finite samples, and adjusting for rounding.
I'm also modelling cellphones and reverberations systems for clients within IRL.
And a number of other things.

I've recently worked out maximum likelihood and M-estimate theory for the model

signal = signal + signal * noise

that is Y_i = m_i(theta) + s_i(theta) * e_i for i=1,...,n

where the "noise" e_1 ... e_N are independent and identically distributed and the location and scale signals are smooth functions of an unknown vector theta.

CLIMATE CHANGE

I've developed estimation theory for mixtures of distributions and applied it to rainfall.
I've also modelled temperature and riverflow series, and applied cusum extensions to detect trend and changing variability in climate series.
Currently I'm modelling the effect of trend and changing variability on extremes (eg extreme temperature).


OTHER INTERESTS.

Kit Withers
c.withers"at"irl.cri.nz (work), kitw"at"slingshot.co.nz (home)
- I use "at" for @ to avoid spam -
phone +64 4 569 0459 (work), +64 4 934 4477 (home), fax +64 4 569 0003 (work),

1