Newsletter 67 Index

Australian &
New Zealand Journal of Statistics

Newsletter Archive

Join the NZSA

Feedback to Editor

New Zealand Statistical Association Newsletter 67

March 2008

Local News

AgResearch
At Lincoln, we recently farewelled David Baird who has left AgResearch in December after 24 years. David has gone to live in Wanaka, where he has set up in private practice under the name VSN (NZ) Limited. We wish him all the best with his new venture. He'll be continuing to spend a lot of his time on GenStat (with VSN International), and is contracting 20% of his time back to AgResearch, so he won't be idle! David's position as biometrician at Lincoln has been advertised but has yet to be filled (any takers?).

At Palmerston North, we welcome Dongwen Luo to AgResearch and our section. Dongwen studied for his PhD at Massey and then Macquarie (Sydney) Universities with Professor Graham Wood on the geometry of generalised linear models. After undertaking a post-doc position in computational bioinformatics at Macquarie, Dongwen returned to New Zealand in 2004 and took up a position as Statistician at the Fonterra Research Centre in Palmerston North. For the past two years he has been in Perth (WA) having a break from statistics and a taste of the business world.

This summer, Aaron Bryant worked as an intern with Roger Littlejohn at Invermay on hidden Markov models with feedback for deer feeding data. Also, Yann Poulouin, a Masters student from France, is on a work placement for six months from February with Neil Cox at Ruakura.

Congratulations to Ken Dodds for receiving a C. Alma Baker Fellowship for travel to the United Kingdom to visit Scottish Agricultural College for approximately three weeks during April to work on the analysis of whole genome marker data for estimating genetic value.

On the conference scene, Ken Dodds attended the 3rd International Congress on Quantitative Genetics in China in August and talked on relatedness: “Using genetic markers in un-pedigreed populations to detect a heritable trait”. Martin Upsdell attended the 5th Biosecurity Summit in Auckland in October. MAF, who hosted the summit, are a key customer for Container Scan, which Martin is working on. David Baird and Dave Saville attended the Biometrics by the Beach conference at Coffs Harbour, NSW in December, and Dave presented a paper entitled “A design for cultivar evaluation trials sown in a serpentine fashion”.

Dave Saville

University of Auckland
The department is delighted to congratulate Chris Wild FRSNZ on his election as a Fellow of the Royal Society of New Zealand in November for his “very significant contributions to both statistical methodology and statistical education.” Chris is our third FRSNZ, with Alastair Scott and George Seber. Congratulations Chris on this richly deserved achievement!

Meanwhile, the mantle of Head of Department passed from Alan Lee to Chris Triggs, who was instituted in a departmental coronation ceremony in February. Coronation regalia befitting of the occasion failed to gain budget approval, but a cardboard Burger King crown was found to suffice. Welcome to the job, Chris, and our thanks to Alan Lee for heading the department through 2007.

Also in the Marsden round, Yong Wang and Sharon Browning both gained Fast Start grants. Yong's grant aims to develop his algorithms for fast computation for fitting mixture models, and Sharon's covers new methodology for determining the genetic contributions to disease.

Our new lecturer Stephane Guindon, appointed last year, has achieved a rare honour: his article with Olivier Gascuel, 'A simple, fast and accurate algorithm to estimate large phylogenies by maximum likelihood', published in Systematic Biology in 2003, has topped 1000 citations! Fewer than 25 scientific articles published since 2003 have been cited 1000 times or more. (How many of these 25 have the word 'simple' in the title remains unknown.) Stephane's algorithm is implemented in the software PhyML. Citations are still rolling in - as I write the total is 1026 and counting...

Three of our completing PhD students have gained prestigious postdoctoral positions around the world. Christian Roever and James Russell successfully defended their theses last year. Christian has now taken up a postdoctoral fellowship at the Max-Planck Institute for Gravitational Physics (Albert Einstein Institute) in Hannover, Germany. James is basking in the tropical sun on Reunion Island in the Indian Ocean, where he is studying the impacts of invasive mammals on seabirds. His ambition is to be the first to find the nesting site of the near-extinct and almost never-seen Mascarene petrel. Steven Miller, who has just submitted his thesis, has taken up a postdoctoral appointment at Trinity College Dublin, studying the movements of ancient peoples through Europe using genetic, linguistic, and archaeological data.

Our current PhD students have also been busy. Lyndon Walker's research on inter-ethnic cohabitation was featured in the New Zealand Herald in December with the intriguing title 'Maori mix up love lives'. Radio interviews on Radio Waatea and Radio PI followed. Particular interest was generated by the decrease in the proportion of Maori people who live with a Maori partner.

Jenny Wilcock won the Early Career prize for best oral presentation at the Biometrics Australasian Regional Conference in Coffs Harbour in December. Also at the conference were invited speakers Russell Millar and Nick Horton, who is visiting the department from Smith College, Massachusetts. Nick has already made stellar contributions to department life, including an excellent talk on statistical education backed up with stunning culinary hospitality. He will not be allowed home if he produces another dinner like his recent welcome for the Longitudinal Data Analysis workshop participants!

Rachel Fewster

University of Otago, Maths and Stats
Two occurrences of note in the Department of Mathematics and Statistics at Otago in February have been Richard Barker taking over as HoD and the arrival of new lecturer Chris Fonnesbeck. Chris is a Canadian by birth and has degrees from University of British Columbia and University of Georgia. In his own words, lifted from the Blogosphere, he is “a marine mammal biologist, statistician and ex-Vancouverite working for the State of Florida's remote research outpost in Atlanta, GA”. Perhaps he will be able to do something about the long absence of dugongs from the waters of Otago Harbour.

Irene Goodwin

Victoria University
As foreshadowed in the last Newsletter of 2007, we have welcomed the arrival of a new Consulting Statistician: Nokuthaba Sibanda. Nokuthaba landed in Wellington in mid September 2007, coming from London, where she did her PhD at Imperial College and later worked as a Research Fellow at the London School of Hygiene and Tropical Medicine. Since arriving, Nokuthaba has settled into the Group well and has been kept pretty busy fielding consulting requests. Also, Nokuthaba's son is now attending the same school as Ivy Liu's twins and John Haywood's daughter; more probable than it first sounds, perhaps, but still an interesting ‘local cluster’.

Everyone is now back from their recent sabbaticals, with John Haywood returning first in December 2007, in time for Helen to start school just as she turned five. Helen thinks school is great, despite (not because of, really!) a long summer break after her first two weeks at school. Richard Arnold came home in mid-February 2008, having spent three months at Waseda University (Tokyo) visiting former colleague Yu Hayakawa, then four months in Seattle at the University of Washington, visiting Jon Wakefield in the Department of Biostatistics. Richard spent this time working on MCMC estimation in capture-recapture modelling. Obviously Richard's work with Yu Hayakawa went well, because at the time of writing Yu is enjoying a return two-week visit to VUW to work jointly with Richard and Stefanka Chukova. The whole Group are excited to see Yu back in Wellington, and we're celebrating with a number of social events while she's here. We're also looking forward to a return visit from Nokuthaba’s predecessor, Colleen Kelly, who will be in Wellington for almost a week in mid March.

Estate Khmaladze also returned to Victoria at the end of January 2008 after a busy time away. On his way to Europe in July 2007 Estate gave two seminar talks in Sydney, to Mathematics and Statistics departments separately. Then in September he was an invited speaker at the 14th Workshop on Stochastic Geometry, Stereology and Image Analysis in Neudietendorf (near Erfurt), Germany. Estate was speaking on the differentiation of set-valued functions, and gained the impression that geometers and image analysts will certainly use such techniques in an enormous variety of applications. In October Estate gave a public presentation to the Georgian Mathematical Society and was also made a Doctor of Business Honoris Causa by the Georgian-American University in Tbilisi. Estate then spent two weeks in Karlsruhe in active discussion with three German professors of geometry and stochastic geometry. Towards the end of his sabbatical, Estate was an invited speaker at the Platinum Jubilee Conference of the Indian Statistical Institute in Kolkata and at a second conference in Hyderabad. He then spent some time in South Indian jungles, riding elephants and living on a boat, before giving a series of linked presentations over three weeks at Hong Kong Baptist University, on Change-set problems, random tessellations, and set-valued analysis.

Although not on sabbatical, Dong Wang recently spent time in both Beijing and Melbourne, returning to VUW at the start of Trimester 1. Similarly, Stefanka Chukova was travelling and visited the US over the NZ summer. Stefanka has great news about her new grand daughter, Kara Whitney Battestilli, who was born on 12 December 2007 and is a very happy, peaceful baby. Stefanka also visited Professor Harry Perros at North Carolina State University, where they made some initial progress on a new joint research project related to service science.

David Vere-Jones has been busy recently, as one of the Organising Committee of the February 2008 Evison Symposium held at VUW to commemorate the contribution of Frank Evison (former Professor of Geophysics at VUW) to earthquake forecasting and related fields. Further details are available on the Evison Symposium website.

David notes that there is some focus on statistical seismology because of its growing role in the development and assessment of time-varying estimates of earthquake risk. In conjunction with the Evison Symposium, a raft of friends and visitors are currently in NZ: Yosi Ogata, Jiancang Zhuang and Takai Iwata, all currently based in Japan, Yan Kagan, David Jackson and others from the USA, Annemarie Christofferson, Stefan Wiemer and others from Europe. In addition, two younger [so DVJ wrote…] visitors are spending three months working with the joint Stats/Geophysics Group at VUW. They are Katerina Orfanogiannaki from Athens (hidden Markov models) and Abigail Jimenez from Spain via Canada (cellular automata and other models for self-organizing criticality). The Evison Symposium is partly supported by the Earthquake Commission (EQC), which is also backing a proposal to set up a testing centre in NZ (one of only three internationally) for evaluating probability forecasts for earthquakes. David also noted that, “finally, Vol2 of Daley and Vere-Jones has actually appeared, and is definitely the LAST of these efforts!” For more details, see: Daley, D.J., Vere-Jones, D., An Introduction to the Theory of Point Processes, Volume II: General Theory and Structure, Springer, 2nd ed. (2008), 573 pages. A bit of light bedtime reading, perhaps…

Finally, there are a number of talks in the Wellington region in the coming weeks, many of which may be (just?) after the Newsletter appears. Here are some details provided by David Harte, the current Wellington Statistics Group (WSG) Convenor. The first three are WSG talks:
4 Mar, 6:00pm: Martin Bland on “Risedronate, the BBC, and me” at the Old Government Building.
17 Mar, 6:00pm: Jiancang Zhuang on “Visualizing patterns in earthquake clusters by using a point process model”.
3 Apr, 6:00pm: Ray Chambers on “Measurement Error in Auxiliary Information”.

Other known talks in Wellington in the next month or so are:
7th Mar, 12:00pm: Martin Bland on “The Tyranny of Power” at Victoria University.
2 Apr, 3:00pm: Ray Chambers on “Robust Prediction of Small Areas Means and Distributions” at Statistics NZ.
4th Apr, 12:00pm: Ray Chambers on “Maximum likelihood under informative sampling” at Victoria University.
7 Apr, 3:00pm: Ray Chambers on “Estimation of the Finite Population Distribution Function” at Te Puni Kokiri.
8 Apr, 1:00pm: Steve Haslett, Geoff Jones and Alasdair Noble on “Small Domain Estimation in NZ Unemployment and Maori Expenditure” at Statistics NZ.
Some further details about all of the above talks can be found at: http://nzsa.rsnz.org/local_groups.shtml.

John Haywood

Wellington Statistics Group
Since the beginning of 2007, there have been nine talks given to the Wellington Statistics Group. These were:

7 Feb 2008, John Lewis, Weta Digital
4 Dec 2007, I-Ming Liu, Victoria University
20 Nov 2007, Mike Camden and Paul Cowie, Statistics NZ
24 Oct 2007, Jim Renwick, NIWA
18 Sep 2007, Peter Thomson, Statistics Research Associates
6 Aug 2007, Nick Longford, SNTL, Reading, England
9 May 2007, Shirley Pledger, Victoria University
18 Apr 2007, Stephen E. Fienberg, Carnegie Mellon University, Pittsburgh
22 Mar 2007, Richard Arnold, Victoria University

Further details of these and all previous talks can be found on the NZSA Local Groups web page. This web page also contains contact details of the group, names of sponsors, and details of forthcoming talks. A link can also be found here so that anybody can add or delete their name from the mailing list.

If anybody is visiting Wellington at a time coinciding with a talk, then you are most welcome to attend. No registration is required. People should not feel reticent about offering a talk that they would like to present. Many work somewhat isolated from other statisticians and often have little opportunity to discuss their work with others. Hopefully we can provide a forum for these people too.

We are grateful to all the WSG sponsors: Victoria University of Wellington, Statistics New Zealand, the Ministry of Social Development and Statistics Research Associates Ltd. Thanks also to Ivy Liu who dealt with the room bookings for 2007, and Alistair Gray who continues to look after the finances and refreshments at each meeting. John Haywood is now back from UCLA and is again helping with room bookings, etc.

David Harte

University of Waikato
It has been a relatively quiet time for all in the Statistics Department at the Waikato University. Although we are still waiting for replacements for James Curran and Nye John, there is good news on that front. We are hoping to appoint a new lecturer very soon, and also that the vacant chair will be filled before the year is out. This is probably just as well as Ray Littler, who has been part-time in the department whilst filling the role of Deputy - sometimes Acting - Dean of the School of Computing and Mathematics Sciences, retired at the end of the year. Fortunately for us, he is continuing to contribute to some of the teaching in the department. Another impending departure is David Whitaker, who plans to retire mid-year. In preparation David has stepped down as Chairperson of the Department and Murray Jorgensen has now taken on that role. At present we are hosting David Johnson from Loughborough, who is also contributing to the teaching in the department.

July 7 and 8, 2008, will see Waikato University host the New Zealand Statistical Association 2008 Conference.

Recent Seminars In The Department:
Russell Millar (University of Auckland) “Assessment of Hierarchical Models for Count Data.”
Rolf Turner (University of New Brunswick) “Direct Maximization of the Likelihood of a Hidden Markov Model”.

Judi McWhirter

Crop & Food Research
In December Bob Burn and Fiona Underwood from Reading University (UK) ran their Practical Bayesian Data Analysis courses at Crop & Food Research in Lincoln and AgResearch Grasslands in Palmerston North. The team who did the Aotearoan organising (Ruth Butler, assisted by Andrew McLachlan, Esther Meenken and John Koolaard from AgResearch) were very pleased by the turn-out - thirty-nine when they had initially been expecting twenty or twenty five - primarily from the CRIs and NZ universities, although one keen person came from RMIT in Melbourne. Bob and Fiona did a great job, despite fighting jet-lag and the after-effects of a heavy term of teaching in the UK. Most of the participants agreed that although they had found the course challenging, it was an excellent introduction to the ideas: everyone thought the course had been very enjoyable and useful.

Duncan Hedderley

Massey University, Albany
At the end of November we officially farewelled Jeff Hunter at a function to mark his retirement. This was a fitting tribute to his work that was attended by many including former colleagues from Palmerston North and Auckland University. Paul Cowperthwait and Graeme Wake edited a special issue of the Journal of Applied Mathematics and Decision Sciences as a tribute to Jeff.

Danny Walsh and Marie Fitch both presented workshops at the NZAMT conference in September.
Howard Edwards attended the 2007 Genetic and Evolutionary Computation Conference (GECCO 2007) at University College London July 7-11 to present a paper and a poster based on joint work with Matthew Walker and Chris Messom. He also attended the Royal Statistical Society's annual conference in York the following week. In the preceding week Howard attended the conference in honour of Emeritus Professor John Deely at the University of Canterbury, Christchurch, as one of John's first PhD students.

Paul Cowperthwait attended the NZ Hydrological Society Conference on “Water and Land” in November where he presented a paper on “Developments of fine scale structure for point process models of rainfall” (co-authored with Valerie Isham and Christian Onof).

Marie Fitch completed her Masters at the end of 2007 and is now beginning a PhD under the supervision of Beatrix Jones and James Curran. Lillian Werner also completed her Masters and is now back in the USA working at the medical centre in Hershey Pensylvania (yes, the place where they make the chocolate!)

Beatrix Jones presented a talk entitled “Using paths through graphical models to dissect pairwise relationships” at the WNAR meeting of the International Biometric Society, June 24-27 at the University of California, Irvine.

Marie Fitch

Fonterra
Since the last report we have recruited Stella Sim, poached from Livestock Improvement (an organisation affiliated to Fonterra). Stella is based at Te Rapa and has been responsible for upgrading the R skills (and to a lesser extent the Bayesian knowledge) of the rest of the team. Roger Kissling will be attending the meeting of the Codex Committee on Methods of Analysis and Sampling, held in Budapest in early March, as technical advisor to the NZ government delegation. Roger and Rob Crawford have been preparing comments on some of the papers to be discussed at that meeting.

In May Rob Crawford will be attending the International Dairy Federation Analytical Week, held this year in Switzerland, specifically to attend the meeting of the Joint Action Team on “Statistics and Sampling” of which he is the chairman.

Roger Kissling

University of Canterbury
Raazesh Sainudiin (Raaz) has recently joined us as a Lecturer in Statistics. Raaz (pictured right) has research interests in statistical inference for stochastic processes embedded within stochastically evolving networks. Examples include statistical decision problems in population genetics, phylogenetics, and ecological genetics. Raaz is also interested in set-valued mathematics and statistics that account for the numerical and/or empirical resolutions of a statistical experiment. Such set-valued experiments are particularly interesting if their index map is highly nonlinear and/or they rely on heuristic, local, point-valued algorithms to implement their decision procedure. For more details see: http://www.math.canterbury.ac.nz/~r.sainudiin/ResStmntRaaz.pdf.

Congratulations to Jennifer Brown on being promoted to Associate Professor. Timothy Robinson, from University of Wyoming, is visiting the department until June 2008. Timothy’s background is in industrial statistics and is working with Jennifer on a number of environmental statistics projects. Jennifer has also been awarded a Geospatial Research Centre grant to investigate the feasibility of using heat-seeking equipment on remote controlled planes to locate pest animals.

Associate Professor Pete Smith (pictured right) has joined the group part-time (jointly with electrical engineering. Among other interests, Pete is interested in stochastic models for wireless communications.

Dominic's masters student Jason Bentley won a prize for best student presentation at the 2007 Spring Bayes Conference in Brisbane for his talk on “Exact Markov Chain Monte Carlo and Bayesian Variable Selection”. This prize continues his successful run after being awarded one of the HRS student prizes at NZSA2007. Well done Jason!

The 12th annual phylogenetics conference, Whitianga '08, was held at a lodge in Whitianga on the Coromandel from 10-15 February 2008. Talks were presented by Mike Steel, Charles Semple, Raaz Sainudin, and Postdoctoral Fellow Bhalchandra Thatte; our PhD students, Klaas Hartmann, Mareike Fischer and Beata Faller; and visitors from Germany, Simone Linz and Tanja Gernhard.

Carl Scarrott

Massey University, Turitea
In early September 2007, Alasdair, Steve, and Geoff attended the SAE2007 conference in Pisa, where between them they gave three talks and took 27 photographs of the leaning tower.

In October Ganes and Alasdair organized another Palmy Stats day, which attracted 48 participants. The keynote speaker, Ian Westbrooke, gave an interesting account of the diverse problems arising from his work at the Department of Conservation.

Martin Hazelton travelled to Dunedin later in October to give a seminar at Otago University (“From Estimation of Traffic Flows to Deconvolution of Densities: Some Statistical Linear Inverse Problems”). In January this year he visited Adrian Baddeley at the University of Western Australia and gave a one-day course on Advanced Regression Modelling under the auspices of the Statistical Society of Australia.
Jonathan Godfrey finally got married to Olivia on 27th January in Wellington. They honeymooned in Cairns. The wedding statistics: 140 guests present, including the first time that all six of his father's kids were in the same room at the same time; 13 of Olivia's mother's kids were present (a record for them too), and seven guide dogs (six working and one retired). The honeymoon statistics are not available. For the record, Jonathan now has an additional six sisters-in-law, and ten brothers-in-law, but only one mother-in-law: that's on top of the ones already accumulated from the Godfrey side.

The New Year brought us a New Institute. Massey’s restructuring of Engineering saw us leave the Institute of Information Sciences and Technology (which now no longer exists) and join the Institute of Fundamental Sciences. Someone should make a board game out of this. We are still in the same offices, although there is a plan to co-locate us with the other Fundamentalists before the next restructuring.

Steve Haslett was a moderator at the Mathematics in Industry Study Group, 28th January - 1st February, in Wollongong, and then stayed on to work with Ray Chambers. Later in February Steve went to the Asian Institute of Technology in Bangkok to assist in the running of a workshop on small-area poverty estimation, as part of a new service agreement with the World Food Programme. Later trips to Cambodia, Bhutan and East Timor are in the pipeline.

Geoff Jones attended the Australasian Region Biometrics Conference in Coffs Harbour, "Biometrics by the Beach" in early December and gave a talk on “Modelling growth patterns via constrained hierarchical b-splines”. He returned just in time to join Alasdair Noble in delivering the first day of a very successful course on “Bayesian Models for Diagnostic Testing”, the main presenters being Wes Johnson and Ian Gardner from the University of California. Geoff then left for a family wedding in Malaysia, and managed to give a seminar on “Measuring poverty and malnutrition” at the University of Malaysia in Sarawak before being whisked off for a sumptuous lunch.

Geoff Jones

Seal-ed with a kiss? Geoff Jones and a mystery guest at the conference dinner in Coffs Harbour.

Scion (NZ Forest Research Institute)
We are now all working under the ‘Scion’ brand since our ‘Ensis’ joint venture with CSIRO has been replaced by a co-operative agreement.
Phil Wilcox, Rod Ball and MapNet (the NZ network of gene mapping researchers) contributors have a concept proposal accepted by FRST for a ‘Virtual Institute of Statistical Genetics’ (VISG). The full proposal has just been submitted. The odds are now 50:50. The proposed VISG consists of statisticians and geneticists from CRIs and Universities and will develop statistical methods to tackle the large amounts of genomic data becoming available with increases in genotyping technologies, and new applications becoming possible for gene mapping across sectors in New Zealand. The VISG will be managed by Phil, with oversight from a governance board including Bruce Weir (U. of Washington), Peter Visscher (QIMR) and Elspeth MacRae (Scion). Statisticians involved include Rod Ball (Scion) , Nihal De Silva (HortResearch), Mik Black (Otago), Sharon Browning (Auckland), Ken Dodds (AgResearch) and Benoit Auvray (AgResearch).
Five projects have been proposed including large datasets (e.g. whole genome mapping and prediction), gene networks (e.g. relationships between expression levels of many genes, proteins and metabolites in relation to biochemical pathways), structural genetics (polyploidy and copy number variation), population structure, experimental design, plus under-pinning methodology (e.g. algorithms, MCMC samplers, convergence etc). The VISG expects to fund several PhD students or PostDocs, who may also be hosted by Bruce Weir at the University of Washington for some time. Expressions of interest to Phil Wilcox or
Rod Ball .
Rod Ball has had a paper: “Quantifying evidence for candidate gene polymorphisms - Bayesian analysis combining sequence-specific and QTL co-location information” published in Genetics, and has been promoted to principal scientist. He is heading to Finland in April to develop collaboration on hybrid MCMC and analytical methods for whole genome models with Mikko Sillanpää at the University of Helsinki on an ISAT travel grant. While there he will give a short course on association mapping and attend the IUFRO conference on modelling and simulation at Koli, Finland (near Joensuu). Olkaa hyvä!

Rod Ball

Return to top