Indian Statistical Institute
Statistics and Mathematics Unit, Bangalore Centre

Celebration of Statistics at ISI, Bangalore

A Conference in Honour of Sunanda Bagchi

January 16, 2020

This event is being organized to celebrate Professor Sunanda Bagchi's career at the Indian Statistical Institute.

Speakers: Bikas Sinha (ISI, Kolkata), Mausumi Bose (ISI, Kolkata) .

Time: 11.30 AM - 12.30 PM and 2.00 PM - 5.30 PM.

Date: January 16, 2020

Venue: SSIU Seminar Hall (II floor)

Programme
Schedule:

 12.30 - 2.00 PM : Lunch Break 2.00 PM - 3.00 PM : Mousumi Bose: Optimal Crossover Designs for Inference on Total Effects 3.00 - 3.30 PM : Tea Break 4.30 - 5.00 PM : Coffee and Snack 5.00 PM : Reminiscences

 Speaker: Bikas K. Sinha         [Top] Title: Some Aspects of Optimal Covariate Designs in Factorial Experiments Abstract: We initiate a study in the context of 2^n -factorial experiments involving the question of optimal allocation of covariate values. There is one controllable quantitative covariate (x) and it is assumed to 'cover' two experimental units at a time. Earlier we dealt with block design set-up [Sinha et al. (2014)]. Here we take up a factorial set-up and address the question of optimal allocation of the covariate values. Results are illustrated for 2^2- and 2^3-factorial experiments. [This is joint work with Professor Prasad Rao, also retired from ISI, Kolkata]

 Speaker: Mousumi Bose         [Top] Title: Optimal Crossover Designs for Inference on Total Effects Abstract: Crossover designs are widely used in many areas, particularly in clinical trials, where each patient receives a sequence of drugs over some time periods. This results in two types of treatment effects, a direct effect and a carryover effect. Several optimality results are available for inferring on these two effects separately. However, a designed experiment recommends a single treatment for use over longer time periods, and so, it is important to study the sum of the direct and carryover effects of the same treatment, that is, the total effect. Not much is known on the optimality of designs for this total effect. In this talk, we focus on this problem and propose universally optimal designs for total effects under a non-circular model and correlated errors. We also report some highly effcient designs in this context. (Joint work with S. Aboukhamseen and S. Huda, Kuwait University)

 Speaker: Bikas K. Sinha         [Top] Title: Social Network Analysis for Understanding of Animal Behavior Abstract: A Social Network is visually represented by a diagram called Network Diagram. Based on such a network, we can explain concepts such as (i) isolated vertices (ii) directed arcs, (iii) symmetric arcs, (iv) cliques, (v) out-degree : reaching out / expansiveness, (vi) in-degree : popularity and (vii) reciprocity : symmetric relation. Degree Centrality, Closeness Centrality, Betweenness Centrality, Eccentricity, sociometric status are also other parameters of relevance. We will closely follow Amelia Coleing [Bioscience Horizons: The International Journal of Student Research, Volume 2, Issue 1, March 2009, Pages 32-43] : The application of social network theory to animal behavior In this presentation we will discuss relevant and natural networks formed by a group of 10 Asian Elephants housed at Chester Zoological Park, UK. The group is composed of seven females and three males. The group has four juveniles

Organized by Stat-Math unit, ISI Bangalore.

Note: No travel support will be provided for attending the event. Please contact jaydeb AT gmail DOT com for any further queries