While doing my masters in Statistics in the early 1980s, I had the good fortune of spending quality time with my teacher Prof. T K S Unnithan, when we waited for the bus from campus back to Trivandrum town. Though I did not get a good grip of, he gave a lot of conceptual insights to some of the classical results in Probability and Statistics, during this waiting. One thing I still remember is the way he associated concepts with human experience sans mathematics. According to him Probability and Statistics is the art and science of learning from experience. Whenever I get a chance to take an introductory class on Probability and Statistics, I use this as defining our discipline. We have many definitions of Statistics, see Gupta and Kapoor (2000) for a compilation from the early days, but none is as striking as the one conceived by my teacher. When I see new generation courses like Statistical Learning, Machine Learning, Data Mining, Artificial Intelligence and the course contents, I am convinced that the way he conceived Probability and Statistics was so comprehensive.
Experience can be in the form of observations/ data and when we need to take a decision using it we need computational power. Rao emphasized both the theory and the methodology developed based on it, as applied to practical problems. His two books on Statistics vouch for this. In the preface to his more influential book, Rao (1965) records: “Under the influence of Fisher and Mahalanobis I have come to appreciate Statistics as the new technology of the present century”. When we realize the explosive power computer gave Statistics by the last decade of the last century, we understand how prophetic Rao’s words were.
Breiman is not just known for his work in Probability or for his classic book Probability (1968) for generations, but also for cutting-edge statistical computation. Even in the 1980’s, he would advise a student considering a career in statistics to: “remember that the great adventure of statistics is in gathering and using data to solve interesting and important real world problems.” Having realized the importance and need for statistical computing early, he established the Statistical Computing Facility at the department of Statistics, University of California, Berkeley in early 1980’s. Now, in the paradigm shift we find in Statistics and in the practice of it, one may notice that he is best known for developing “Random Forests” and “Regression Trees”.
Probability and Statistics has come a long way from the early 1980’s, in the way it is taught and practiced. Kerala Statistical Association has tried to move along with this paradigm shift in their deliberations over the years. I wish the Kerala Statistical Association the very best in its service to statistics fraternity.
Breiman, L (1968), Probability, Addison-Wesley Publishing Company, Inc., Reading, Massachusetts.
Gupta, S C and Kapoor, V K (2000), Fundamentals of Mathematical Statistics, Sultan Chand and Sons Private Limited, New Delhi.
Rao, C R (1965), Linear Statistical Inference and Its Applications, John Wiley and Sons, New York.
