SOP(s) from 1998

I still fondly recollect the ecstasy I experienced after learning how to compute the center of gravity of a lamina using calculus. I kept on cutting paper pieces, finding their centers of gravity , and balancing them on match sticks. Neither the computation nor the balancing alone could give me the satisfaction that I derived by doing both simultaneously.

This childish enthusiasm gradually crystallized into a more mature form as I stepped into my higher secondary study in statistics. Initially I did not have much fascination for the subject until one small experiment changed my view. It was the age old experiment of tossing a coin. I tossed a coin many times and plotted the proportions of heads. As I watched the resulting zigzag curve converging to one point, the notion of probability suddenly became a tangible object to me. I felt it as a law which even the most trivial thing must obey. Statistics, thus, became to me a study that is intimately related with the workings of nature.

The next five years saw me engrossed deeply in the study of this fascinating subject in the Indian Statistical Institute. These were my B. Stat. (Hons.) and M. Stat. years. Here I met another turning point of my academic career --- the computer. The idea of ``something that works'' seemed to reach a culmination with this machine. The enormous interaction between computers and statistics enhanced my liking for both the subjects. My self study on computers helped me to explore topics like finite automata and compilers. I had a course on FORTRAN , but learned C by myself. My achievements along this line include writing a graphics package (which I wrote right from the scratch). This package can draw 3 dimensional graphs. I found great pleasure in searching out all the necessary algorithms from the literature. Adapting one hidden-line-removal algorithm to accommodate some extra positions of the viewer was really challenging. The package helped all my classmates to make plots used in our statistical courses. Another program which I am proud of writing is a parser generator. This program-writing program is only slightly less powerful than the standard parser generator YACC for performing small tasks.

Side by side with these amateurish pursuits, I steadily received highly rigorous training of statistics and mathematics. I had frequent opportunities of seeing statistics work in practice --- not only by taking part in some statistical counsellings in our institute, but also by doing active project works. My participation in the statistical counsellings revealed to me many mistaken notions that common men (and even scientists from other fields) harbored about statistics. Once we had great trouble in explaining to a gentleman that it was not too wise a thing to try to fit a model with 72 parameters to a dataset of size 26 only! Such experiences have led me to believe that it is important to introduce statistical methods at a more popular level, and have strengthened my aspiration to become a teacher.

Among my project works an analysis of pollution data collected in Calcutta features foremost. The data --- a time series data with missing observations --- had some interesting features. For instance, I found a high correlation among $NO_x$ and $SO_2$ quite inexplicable, and had to visit the chemical laboratory to investigate the data collection procedure. In another of my projects I developed a new method for simulating data from spatial Poisson processes.

An important contribution to my studies has been made by my class presentations. They used to be one hour or ,mostly, two hour lectures on topics related to mathematics and statistics. Secretary problem, distance sampling, Wold decomposition, ridge regression were some of the statistical topics, while the important mathematical topics included complex measure and algebraic codes. The latter talk --- a two and a half hour lecture --- was one of the talks that I liked most. It started with an introduction to error correcting codes and went upto one efficient algorithm for decoding BCH codes. These talks have helped me to gain proficiency in expressing myself orally. This gift is quite vital for teaching --- the career which I plan to take up in future. By the way,I have also penned an article called {\bf `The Secretary Problem: Optimal Stopping'} in the Resonance (Sep, 1996, Vol 1, No. 9) , a journal published by the Indian Academy of Sciences.

During the five years that I spent as a student in the Indian Statistical Institute (Calcutta) I was always among the top three in my class (except in one semester when I stood fourth). I enjoyed full government stipend and contingency grants during these five years. Also, I received nine cash awards for good performance in my class.

Charming as my close contact with statistics was, it also revealed to me the seamy side of the subject. Its rampant misuse and wrong interpretation baffled me. For a time I was frustrated --- as deeply as I had loved the subject earlier. But, with the frustration subsiding with time, I began to think more optimistically. I wanted to face the situation actively, and my standpoint towards the subject changed form that of a fascinated onlooker to that of an active researcher.

Side by side with my enthusiasm for applied statistics, I nurtured a liking for the abstract language of mathematics --- especially algebra. Indeed, a talk on Gr\"{o}bner bases has brought me the P C Mahalanobis International Symposium in Statistics Award (1997-98). This award is given annually to ``the most outstanding M. Stat. student''. The selection is based on a nomination followed by a one hour lecture presentation. In the meantime, after finishing my M. Stat. degree, I had planned to apply abroad. But some domestic problems thwarted my plans, and I temporarily devoted myself to a study of DNA sequencing. This investigation has revealed some interesting results. I hope to summarise my findings in print by May, 1999. Besides this, I am also going to officially teach time series to M. Stat. first year students of the Indian Statistical Institute (Calcutta) in the coming semester. I have also given four lectures to B. Stat.(Hons.) third year students of the same institute on loglinear models. The domestic problems that I mentioned earlier being now over, I hope to join a US university for graduate study --- so that I may get an international exposure, and may enrich my research interests through the contact of a vast spectrum of subjects.

Thank you for considering my application.

ARNAB CHAKRABORTY