Risk Assessment of Future Antibiotic Resistance—Eliciting and Modelling Probabilistic Dependencies Between Multivariate Uncertainties of Bug-Drug Combinations

The increasing impact of antibacterial resistance concerns various stakeholders, including clinicians, researchers and decision-makers in the pharmaceutical industry, and healthcare policy-makers. In particular, possible multidrug resistance of bacteria poses complex challenges for healthcare risk assessments and for pharmaceutical companies’ willingness to invest in research and development (R&D). Neglecting dependencies between uncertainties of future resistance rates can severely underestimate the systemic risk for certain bug-drug combinations. In this paper, we model the dependencies between several important bug-drug combinations’ resistance rates that are of interest for the United Kingdom probabilistically through copulas. As a commonly encountered challenge in probabilistic dependence modelling is the lack of relevant historical data to quantify a model, we present a method for eliciting dependence information from experts in a formal and structured manner. It aims at providing transparency and robustness of the elicitation results while also mitigating common cognitive fallacies of dependence assessments. Methodological robustness is of particular importance whenever elicitation results are used in complex decisions such as prioritising investments of antibiotics R&D.

Vine-Copula-Based Quantile Regression for Cascade Reservoirs Management

This paper features an application of Regular Vine (R-vine) copulas, a recently developed statistical tool to assess composite risk. Copula-based dependence modelling is a popular tool in conditional risk assessment, but is usually applied to pairs of variables. By contrast, Vine copulas provide greater flexibility and permit the modelling of complex dependency patterns using a wide variety of bivariate copulas which may be arranged and analysed in a tree structure to explore multiple dependencies. This study emphasises the use of R-vine copulas in an analysis of the co-dependencies of five reservoirs in the cascade of the Saint-John River basin in Eastern Canada. The developed R-vine copulas lead to the joint and conditional return periods of maximum volumes, for hydrologic design and cascade reservoir management in the basin. The main attraction of this approach to risk modelling is the flexibility in the choice of distributions used to model heavy-tailed marginals and co-dependencies.

VEGETATED HYDRODYNAMIC SYSTEM: PARAMETERIZATION AND STOCHASTIC DEPENDENCE MODELLING

Vegetation as a nature-based solution for increasing flood risk has convincingly shown potential for flood hazard (wave load) reduction but lacks generalized results. In this study we have introduced stochastic dependence modeling using non-parametric Bayesian networks (NPBN) for vegetated coastal systems where the system was parametrized using continuous marginal distributions, and likely (conditional) correlations among variables. The model represented a consistent joint probability distribution and hence can be used to generate physically realistic conditions in data-scare environments. It adds value to numerical modeling by reducing the number of simulations required to get meaningful generalized results. Main findings, that were derived by using a NPBN, help to pave way for implementation of nature-based solutions for a range of realistic conditions that can be found across global coastal foreshores.Recorded Presentation from the vICCE (YouTube Link): https://youtu.be/T6TP0DH0qMw

Utilization Copula in Determination of Shallot Insurance Premium Based on Regional Harvest Results

Shallot is one of the highest-yielding horticultural crops in Indonesia and has the tendency to increase the profits of farmers in Indonesia. But until now in Indonesia there is no insurance for horticultural crops other than corn, whereas the shallot farmers face various sources of risk such as weather changes, pest attacks, or other technical factors that ultimately lead to uncertainty of agricultural yields (revenue risk). To overcome this loss, insurance companies can make products based on shallot yields and shallot market prices. Therefore it is essential to grasp the distribution of risk variables (shallot yields and shallot market prices) that interact simultaneously, not separate from one another. Omitting dependencies among risk variables can cause biased risk estimation. Copula can model the non-linear dependencies and can identify the structure of the dependencies between variables. The suitable copula for modeling yield and price risk of shallot is simulated to compute the premium. Result show that clayton copula is suitable for dependence modelling between risk variables.

On unbalanced data and common shock models in stochastic loss reserving

Introducing common shocks is a popular dependence modelling approach, with some recent applications in loss reserving. The main advantage of this approach is the ability to capture structural dependence coming from known relationships. In addition, it helps with the parsimonious construction of correlation matrices of large dimensions. However, complications arise in the presence of “unbalanced data”, that is, when (expected) magnitude of observations over a single triangle, or between triangles, can vary substantially. Specifically, if a single common shock is applied to all of these cells, it can contribute insignificantly to the larger values and/or swamp the smaller ones, unless careful adjustments are made. This problem is further complicated in applications involving negative claim amounts. In this paper, we address this problem in the loss reserving context using a common shock Tweedie approach for unbalanced data. We show that the solution not only provides a much better balance of the common shock proportions relative to the unbalanced data, but it is also parsimonious. Finally, the common shock Tweedie model also provides distributional tractability.

Recurrent Dirichlet Belief Networks for interpretable Dynamic Relational Data Modelling

The Dirichlet Belief Network~(DirBN) has been recently proposed as a promising approach in learning interpretable deep latent representations for objects. In this work, we leverage its interpretable modelling architecture and propose a deep dynamic probabilistic framework -- the Recurrent Dirichlet Belief Network~(Recurrent-DBN) -- to study interpretable hidden structures from dynamic relational data. The proposed Recurrent-DBN has the following merits: (1) it infers interpretable and organised hierarchical latent structures for objects within and across time steps; (2) it enables recurrent long-term temporal dependence modelling, which outperforms the one-order Markov descriptions in most of the dynamic probabilistic frameworks; (3) the computational cost scales to the number of positive links only. In addition, we develop a new inference strategy, which first upward-and-backward propagates latent counts and then downward-and-forward samples variables, to enable efficient Gibbs sampling for the Recurrent-DBN. We apply the Recurrent-DBN to dynamic relational data problems. The extensive experiment results on real-world data validate the advantages of the Recurrent-DBN over the state-of-the-art models in interpretable latent structure discovery and improved link prediction performance.