Staff profile

Dr Paul Kabaila

Reader in Statistics

Faculty of Science, Technology and Engineering

School of Engineering and Mathematical Sciences
Department of Mathematics and Statistics

Physical Sciences 2, Room 225, Melbourne (Bundoora)

 

Qualifications

BSc, BE (Hons 1), PhD.

Membership of professional associations

Elected member of the International Statistical Institute (ISI), member of the Institute of Mathematical Statistics, member of the Bernoulli Society, member of the Statistical Society of Australia and member of the International Biometric Society, Australasian Region.

Area of study

Mathematics and Statistics

Brief profile

After completing my PhD on the statistical analysis of time series, I worked as a statistical consultant at CSIRO, Division of Mathematics and Statistics for four years. Then I worked as a consultant in applied stochastic processes at Telecom Research Laboratories for two and a half years. I then joined the Department of Statistics at La Trobe University.

At La Trobe, I have taught a very wide variety of statistics subjects, at all levels. In 2009, I wrote the new 4th year subject Theory of Statistics, which I teach across the Access Grid. In 2011, I wrote the new subject HCS4RSP (Research Methods and Statistics for Speech Pathology).

I have held visiting appointments at the Australian National University (Jan - Dec 1991), Oxford University (Mar 1997 - Jan 1998), Monash University (Mar - May and Jul - Sept 2002) and the Australian Mathematical Sciences Institute (Feb - Jul 2008 and Sept 2008 - Feb 2009).

I was awarded an ARC Discovery Grant of $155,000 for the project entitled 'New and computationally feasible methods of constructing efficient and exact confidence limits from count data' (2002-2005, jointly with Chris J. Lloyd). This project has led to important advances in the statistical analysis of count data in fields such as epidemiology, reliability, toxicology and finance.

In 2005, La Trobe was the first university to have its Statistics Degree Program accredited by the Statistical Society of Australia Inc.. I was the Responsible Officer for the application that resulted in this accreditation and two subsequent re-accreditations. Currently, the Statistics Program at La Trobe is the only Victorian undergraduate Statistics Program accredited by the Statistical Society of Australia Inc..

La Trobe, Monash and RMIT universities share the teaching of 4th year statistics subjects through an arrangement known as the Key Centre for Statistical Sciences (KCSS). I am currently the Director of the KCSS.

Current research

My research is currently focussed on (a) the effect of preliminary statistical model selection on confidence intervals and prediction intervals and (b) frequentist confidence regions and prediction regions that utilize uncertain prior information.

I presented a one hour invited talk entitled "Confidence intervals in regression utilizing prior information" at the Workshop on Current Trends and Challenges in Model Selection and Related Areas, held at the University of Vienna, Austria, 24 July - 26 July 2008. I was also the organiser (with Hannes Leeb) of the Special Topic Session on Inferential tools based on model selection, shrinkage and related methods at the 58th World Statistics Congress of the International Statistical Institute held in Dublin, Ireland from 21 to 26 August 2011.

I am currently supervising postgraduate students Dilshani Tissera, Waruni Abeysekera, Sameera Dharmaratne and Yining Kong in these fields. I am also supervising an Honours in Statistics student and a Master of Statistical Science student.

Research interests

Theory of Statistical Inference

- Computationally feasible methods of constructing efficient and exact confidence intervals from count data.

- Frequentist confidence regions and prediction regions that utilise uncertain prior information.

- The effect of preliminary statistical model selection on confidence intervals and prediction intervals.

Time series analysis

- Time series prediction intervals that take account of parameter estimation errors and volatility clustering.

Teaching units

STA3AS Applied Statistics, 3rd year unit, semester 2

STA4SI Statistical Inference, 4th year unit, semester 2

Recent publications

  • Kabaila, P. and Syuhada, K. (2008). Improved prediction limits for AR(p) and ARCH(p) processes. Journal of Time Series Analysis, 29, 213-223.

  • Kabaila, P. (2008) Statistical properties of exact confidence intervals from discrete data using studentized test statistics. Statistics & Probability Letters, 78, 720-727.

  • Giri, K. and Kabaila, P. (2008), The coverage probability of confidence intervals in 2r factorial experiments after preliminary hypothesis testing. Australian & New Zealand Journal of Statistics, 50, 69-79.

  • Farchione, D. and Kabaila, P. (2008). Confidence intervals for the normal mean utilizing prior information. Statistics & Probability Letters, 78, 1094-1100.

  • Kabaila, P. and Tuck, J. (2008). Confidence intervals utilizing prior information in the Behrens-Fisher problem. Australian & New Zealand Journal of Statistics, 50, 309-328.

  • Kabaila, P. and Giri, K. (2009). Upper bounds on the minimum coverage probability of confidence intervals in regression after model selection. Australian & New Zealand Journal of Statistics, 51, 271-287.

  • Kabaila, P. and Giri, K. (2009). Large-sample confidence intervals for the treatment difference in a two-period crossover trial, utilizing prior information. Statistics & Probability Letters, 79, 652-658.

  • Kabaila, P. and Giri, K. (2009). Confidence intervals in regression utilizing uncertain prior information. Journal of Statistical Planning and Inference, 139, 3419-3429.
  • Kabaila, P. (2009). The coverage properties of confidence regions after model selection. International Statistical Review, 77, 405-414.
  • Kabaila, P., Giri, K. and Leeb, H. (2010). Admissibility of the usual confidence interval in linear regression. Electronic Journal of Statistics, 4, 300-312.
  • Kabaila, P. and Syuhada, K. (2010). The asymptotic efficiency of improved prediction intervals. Statistics & Probability Letters, 80, 1348-1353.
  • Lloyd, C. and Kabaila, P. (2010). Letter to the Editor: Some comments on On construction of the smallest one-sided confidence interval for the difference of two proportions, Ann. Statist. 38 (2010) 1227-1243. The Annals of Statistics, 38, 3840-3841.
  • Kabaila, P. (2011). Admissibility of the usual confidence interval for the normal mean. Statistics & Probability Letters, 81, 352-359.
  • Kabaila, P.  and Farchione, D. (2012). The minimum coverage probability of confidence intervals in regression after a preliminary F test. Journal of Statistical Planning and Inference, 142, 956-964.
  • Kabaila, P. and Vicendese, M. (2012). The performance of a two-stage analysis of ABAB/BABA crossover trials. Biometrical Journal, 54, 361-369.
  • Kabaila, P. and Tissera, D. (2012). Effect of a preliminary test of homogeneity of stratum-specific odds ratios on their confidence intervals. Electronic Journal of Statistics, 6, 672-685.
  • Kabaila, P. (2012). Review of Introduction to the Theory of Statistical Inference, by Hannelore Liero and Silvelyn Zwanzig. Boca Raton, FL. CRC Press, 2012. Australian & New Zealand Journal of Statistics, 54, 394-395.
  • Farchione, D. and Kabaila, P. (2012). Confidence intervals in regression centred on the SCAD estimator. Statistics & Probability Letters, 82, 1953-1960. 
  • Kabaila, P. and Giri, K. (2013). Simultaneous confidence intervals for the population cell means, for two-by-two factorial data, that utilize uncertain prior information. To appear in Communications in Statistics - Theory and Methods.
  • Kabaila, P. (2013).  Note on a paradox in decision-theoretic interval estimation. Statistics & Probability Letters, 83, 123-126.

 

Research projects

  • Frequentist confidence regions and prediction regions that utilise uncertain prior information.
  • The effect of preliminary model selection on confidence intervals.
  • Computationally feasible methods for constructing efficient and exact confidence intervals from count data.
  • Time series prediction intervals that take account of parameter estimation errors and volatility clustering.