I was originally drawn to the discipline of Statistics by its
wide applicability. I consider Statistics as a part of Information Science. I
have done research in
robust statistical procedures, saddlepoint approximation for densities,
modality of gaussian mixtures, and ensembles of nomparametric classifiers.
Since 2002, I am doing research
in data mining, particularly in
data preprocessing techniques such as feature selection, instance selection, handling of missing values,
outlier detection, noise reduction, and visualization. I am also interesed in algorithms for data streams and functional data analysis.
A third area of interest is application of distributed and parallel computation to knowledge discovery in massive datasets through the use of
Hadoop and Mapreduce.
LINKS TO FORMER STUDENTS
Last updated December 10, 2011