Sanjay Chaudhuri

Estimation of a covariance matrix with zeros

We consider estimation of the covariance matrix of a multivariate random vector under the constraint that certain covariances are zero. We first present an algorithm, which we call iterative conditional fitting, for computing the maximum likelihood estimate of the constrained covariance matrix, under the assumption of multivariate normality. In contrast to previous approaches, this algorithm has guaranteed convergence properties. Dropping the assumption of multivariate normality, we show how to estimate the covariance matrix in an empirical likelihood approach. These approaches are then compared via simulation and on an example of gene expression. Copyright 2007, Oxford University Press.

Harvesting as a disease control measure in an eco-epidemiological system - a theoretical study

Epidemiology and ecology are traditionally treated as independent research areas, but there are many commonalities between these two fields. It is frequently observed in nature that the former has an encroachment into the later and changes the system dynamics significantly. In population ecology, in particular, the predator-prey interaction in presence of parasites can produce more complex dynamics including switching of stability, extinction and oscillations. On the other hand, harvesting practices may play a crucial role in a host-parasite system. Reasonable harvesting can remove a parasite, in principle, from their host. In this paper, we study theoretically the role of harvesting in a predator-prey-parasite system. Our study shows that, using impulsive harvesting effort as control parameter, it is not only possible to control the cyclic behavior of the system populations leading to the persistence of all species, but other desired stable equilibrium including disease-free can also be obtained.

Toxic phytoplankton-induced spatiotemporal patterns

Here we consider a reaction diffusion system of three plankton populations, a zooplankton feeding on two phytoplankton populations, in two different settings. Firstly, the two phytoplanktons are both non-toxic and both enhance the growth of the grazing zooplankton. Secondly, we assume that one of the phytoplankton releases toxin and thereby inhibits the growth of the zooplankton. Our analytic and numerical study shows that the spatiotemporal distribution of the plankton species is uniform when both phytoplankton populations are non-toxic. However, in the presence of toxin-producing phytoplankton, the biomass distribution of all the plankton populations becomes inhomogeneous.

Qualitative inequalities for squared partial correlations of a Gaussian random vector

We describe various sets of conditional independence relationships, sufficient for qualitatively comparing non-vanishing squared partial correlations of a Gaussian random vector. These sufficient conditions are satisfied by several graphical Markov models. Rules for comparing degree of association among the vertices of such Gaussian graphical models are also developed. We apply these rules to compare conditional dependencies on Gaussian trees. In particular for trees, we show that such dependence can be completely characterised by the length of the paths joining the dependent vertices to each other and to the vertices conditioned on. We also apply our results to postulate rules for model selection for polytree models. Our rules apply to mutual information of Gaussian random vectors as well.

Phytoplankton-zooplankton dynamics in the 'presence' or 'absence' of toxic phytoplankton

The inhibitory effects of toxin-producing phytoplankton (TPP) on zooplankton modulate the dynamics of marine plankton. In this article, we employ simple mathematical models to compare theoretically the dynamics of phytoplankton-zooplankton interaction in situations where the TPP are present with those where TPP are absent. We consider two sets of three-component interaction models: one that does not include the effect of TPP and the other that does. The negative effects of TPP on zooplankton is described by a non-linear interaction term. Extensive theoretical analyses of the models have been performed to understand the qualitative behaviour of the model systems around every possible equilibria. The results of local-stability analysis and numerical simulations demonstrate that the two model-systems differ qualitatively with regard to oscillations and stability. The model system that does not include TPP is asymptotically stable around the coexisting equilibria, whereas, the system that includes TPP oscillates for a range of parametric values associated with toxin-inhibition rate and competition coefficients. Our analysis suggests that the qualitative dynamics of the plankton-zooplankton interactions are very likely to alter due to the presence of TPP species, and therefore the effects of TPP should be considered carefully while modelling plankton dynamics.

The role of reversals in order-restricted inference

A statistical model is said to be an order-restricted statistical model when its parameter takes its values in a closed convex cone C of the Euclidean space. In recent years, order-restricted likelihood ratio tests and maximum likelihood estimators have been criticized on the grounds that they may violate a cone order monotonicity (COM) property, and hence reverse the cone order induced by C. The authors argue here that these reversals occur only in the case that C is an obtuse cone, and that in this case COM is an inappropriate requirement for likelihood-based estimates and tests. They conclude that these procedures thus remain perfectly reasonable procedures for order-restricted inference.

Reversing the Stein Effect

The Reverse Stein Effect is identified and illustrated: A statistician who shrinks his/her data toward a point chosen without reliable knowledge about the underlying value of the parameter to be estimated but based instead upon the observed data will not be protected by the minimax property of shrinkage estimators such as that of James and Stein, but instead will likely incur a greater error than if shrinkage were not used.

Quasiperiodic waves at the onset of zero-Prandtl-number convection with rotation.

We show the possibility of temporally quasiperiodic waves at the onset of thermal convection in a thin horizontal layer of slowly rotating zero-Prandtl-number Boussinesq fluid confined between stress-free conducting boundaries. Two independent frequencies emerge due to an interaction between straight rolls and waves along these rolls in the presence of Coriolis force, if the Taylor number is raised above a critical value. Constructing a dynamical system for the hydrodynamical problem, the competition between the interacting instabilities is analyzed. The forward bifurcation from the conductive state is self-tuned.

Edge selection for undirected graphs

ABSTRACT This article explores an ‘Edge Selection’ procedure to fit an undirected graph to a given data set. Undirected graphs are routinely used to represent, model and analyse associative relationships among the entities on a social, biological or genetic network. Our proposed method combines the computational efficiency of least angle regression and at the same time ensures symmetry of the selected adjacency matrix. Various local and global properties of the edge selection path are explored analytically. In particular, a suitable parameter that controls the amount of shrinkage is identified and we consider several cross-validation techniques to choose an accurate predictive model on the path. The proposed method is illustrated with a detailed simulation study involving models with various levels of sparsity and variability in the nodal degree distributions. Finally, our method is used to select undirected graphs from various real data sets. We employ it for identifying the regulatory network of isoprenoid pathways from a gene-expression data and also to identify genetic network from a high-dimensional breast cancer study data.

Variance estimation for tree-order restricted models

In this article, we discuss the estimation of the common variance of several normal populations with tree-order restricted means. We discuss the asymptotic properties of the maximum-likelihood estimator (MLE) of the variance as the number of populations tends to infinity. We consider several cases of various orders of the sample sizes and show that the MLE of the variance may or may not be consistent or be asymptotically normal.