Research Areas

Statistical Computing: My main interests are in computational algorithms for linear and nonlinear models, optimization algorithms and random number generation. Investigating methods for accelerating the EM algorithm is a current research problem.

Linear Models: I have been a leading contributor to the literature in the area of analysis of nonreplicated experiments over the last 15 years. Main work has been in testing for partial additivity in two-way tables and extensions of the multiplicative models to multi-way structures. I have also done some work on testing for outliers in regression models.

Educational Graphics: Several of my students and I have used modern computational environments to develop graphical tools useful for data analysis and teaching. We primarily relied on Lisp-Stat for developing software for these purposes. I am currently working on making these work as Java applets on the Web.


Selected Publications

Marasinghe, Mervyn G., Duckworth, William D., and Shin, Tae-Sung (2004). "Tools for Teaching Regression Concepts Using Dynamic Graphics,"  Journal of Statistical Education,Volume 12, Number 2 (2004)

Iversen, Philip W. and Marasinghe, Mervyn G. (1997). "Teaching Experimental Design and Analysis Concepts Using Dynamic Graphics,"The American Statistician. 55, 341-351

Marasinghe, Mervyn G., Meeker, William Q., Cook, Dianne and Shin, Tae-Sung (1996). "Using Graphics and Simulation to Teach Statistical Concepts,"  The American Statistician. 50, 342-351.

Marasinghe, M. G. and Boik, R. J. (1993) "A Three-degree of Freedom Test of Additivity in Three-Way Classifications,"Computational Statistics and Data Analysis, 16, 47--61.

Marasinghe, M. G.(1985) "A Multistage Procedure for Detecting Several Outliers in Linear Regression,'' Technometrics, 27, 395--399

Marasinghe, M. G. and Johnson, D. E. (1983) "A Test of Incomplete Additivity in the Multiplicative Interaction Model," Journal of the American Statistical Association, 77, 869-877.


Selected Work-in-Progress


Shin, Tae-Sung and Marasinghe, Mervyn G. "Development of Efficient Maximum Likelihood Algorithms" under preparation. 

Guo, Y, Marasinghe, Mervyn G.,  and Palmer, Reid G. (2004). "Evaluation of Chinese Soybean Germplasm to Improve Soybean Production and Utilization," being revised  for pulication in  Crop Science.

Iversen, Philip W. and Marasinghe, Mervyn G. (2004) "Visualizing Experimental Designs for Balanced ANOVA Models Using LISP-STAT," submitted to the Journal of Statistical Software.

Marasinghe, Mervyn G and Cifuentes, Julio, C.A.(2004). "New Cojugate Gradient Acceleration Method for the EM Algorithm" , under preparation.



 

    Book Reviews

Marasinghe, Mervyn G.  Review of " Numerical Methods in Statistics," by John F. Monahan, Journal of American Statistical Association,
 



Return to my homepage .