To Department of Mathematics and Statistics homepage

Welcome to my home page !



Last modified: Dec 2021


Chaitanya Joshi
Chaitanya Joshi

Ph.D (2011 - TCD), M.Sc. (2003 - IIT-K), B.Sc. (2001 - Mumbai)

Senior Lecturer,
Department of Mathematics and Statistics,
Univerisity of Waikato, Hamilton, New Zealand.
email: cjoshi@waikato.ac.nz

Current Research Interests: Bayesian Prior Elicitation and Robustness, Adversarial Risk Analysis, Bayesian Deep Learning, Modelling in Security and Crime Science


My Research: I am interested in both: (i) Theoretical and methodological research in Bayesian methods as well as (ii) in modeling complex real life processes using statistical methods.

Theoretical and methodological research:
  1. I am interested in challenges around elicitation of subjective priors and the robustness of the Bayesian analysis to the choice of prior distributions. I've recently done some work on developing new classes of prior distributions with Fabrizio Ruggeri.
  2. Going forward, I am hoping to work on developing framework for incorporating subjective expert priors in Bayesian deep learning.
  3. Along with Stephen Joe and Ph.D. student Muhammad Ejaz, I have been developing Adversarial Risk Analysis models for Auctions.
  4. Previous work includes developing computationally efficient methods for Bayesian inference. Along with Stephen Joe and Ph.D. student Paul Brown, we have developed a computationally efficient algorithm based on low discrepancy sequences which can be used in grid-based Bayesian methods such as INLA in collaboration with Havard Rue.
Statistical Modelling:
  1. I am part of the newly formed NZ Institute of Security and Crime Science. I have done some work on modelling crime in collaboration with NZ Police. I am working on building Adversarial Risk Analysis (ARA) models in collaboration with David Rios (ICMAT) and Jesus Rios. I am also involved in modelling and analysis of Family violence data with Devon Polaschek and PhD student Apriel Joliffe-Simpson.
  2. I am working on a project to model and predict traffic flows for Hamilton city using space-time models and deep learning with Masters student Dale Townsend.
  3. From time to time, I collaborate with scientists on interesting modelling problems. This includes recent work with Bruce Clarkson and PhD student Elizabeth Eliot-Hogg on the capacity of urban restored sites to support native birds. Previously, work with Louis Schipper's group on incorporating uncertainty using Bayesian methods. Earlier, I also worked with Daniel Laughlin (now at Univ. of Wyoming, U.S.) on problems related to modeling species distribution. We developed a novel mathematical framework called 'Traitspace' which incorporates the various processes/factors which govern the assembly of ecological communities via their functional traits and predicts the community assembly by using the observed trait values.
Previous work/interests:
  1. Efficient Bayesian inference on Stochastic Differential Equation (SDE) models, Epidemiology, Clinical trials, Demand estimation and Market research.
  2. From 2003 until 2007, I worked as a statistician for a number of leading corporations in the pharmaceutical and market research area.

Research STUDENTS:




PUBLICATIONS

2022