Bayesian network-based spatial predictive modelling reveals COVID-19 transmission dynamics in Eswatini

Modelling COVID-19 transmission at live events and public gatherings is essential to evaluate and control the probability of subsequent outbreaks. Model estimates can be used to inform event organizers about the possibility of super-spreading and the predicted efficacy of safety protocols, as well as to communicate to participants their personalised risk so that they may choose whether to attend. Yet, despite the fast-growing body of literature on COVID transmission dynamics, current risk models either neglect contextual information on vaccination rates or disease prevalence or do not attempt to quantitatively model transmission, thus limiting their potential to provide insightful estimates. This paper attempts to bridge this gap by providing informative risk metrics for live public events, along with a measure of their associated uncertainty. Starting with a thorough review of the literature and building upon existing models, our approach ties together three main components: (a) reliable modelling of the number of infectious cases at the time of the event, (b) evaluation of the efficiency of pre-event screening and risk mitigation protocols, and (c) modelling the transmission dynamics during the event. We demonstrate how uncertainty in the input parameters can be included in the model using Monte Carlo simulations. We discuss the underlying assumptions and limitations of our approach and implications for policy around live events management.

Download Full-text

Review on Predictive Modelling Techniques for Identifying Students at Risk in University Environment

MATEC Web of Conferences ◽

10.1051/matecconf/201925503002 ◽

2019 ◽

Vol 255 ◽

pp. 03002

Author(s):

Mat Yaacob Nik Nurul Hafzan ◽

Deris Safaai ◽

Mat Asiah ◽

Mohamad Mohd Saberi ◽

Safaai Siti Syuhaida

Keyword(s):

Higher Education ◽

Machine Learning ◽

At Risk ◽

Bayesian Network ◽

Predictive Analytics ◽

Dynamic Bayesian Network ◽

Predictive Modelling ◽

Support Vector ◽

Modelling Techniques ◽

At Risk Student

Predictive analytics including statistical techniques, predictive modelling, machine learning, and data mining that analyse current and historical facts to make predictions about future or otherwise unknown events. Higher education institutions nowadays are under increasing pressure to respond to national and global economic, political and social changes such as the growing need to increase the proportion of students in certain disciplines, embedding workplace graduate attributes and ensuring that the quality of learning programs are both nationally and globally relevant. However, in higher education institution, there are significant numbers of students that stop their studies before graduation, especially for undergraduate students. Problem related to stopping out student and late or not graduating student can be improved by applying analytics. Using analytics, administrators, instructors and student can predict what will happen in future. Administrator and instructors can decide suitable intervention programs for at-risk students and before students decide to leave their study. Many different machine learning techniques have been implemented for predictive modelling in the past including decision tree, k-nearest neighbour, random forest, neural network, support vector machine, naïve Bayesian and a few others. A few attempts have been made to use Bayesian network and dynamic Bayesian network as modelling techniques for predicting at- risk student but a few challenges need to be resolved. The motivation for using dynamic Bayesian network is that it is robust to incomplete data and it provides opportunities for handling changing and dynamic environment. The trends and directions of research on prediction and identifying at-risk student are developing prediction model that can provide as early as possible alert to administrators, predictive model that handle dynamic and changing environment and the model that provide real-time prediction.

Download Full-text

A Predictive Modelling Framework for COVID-19 Transmission to Inform the Management of Mass Events (Preprint)

10.2196/preprints.30648 ◽

2021 ◽

Author(s):

Claire Donnat

Keyword(s):

Prediction Interval ◽

Contextual Information ◽

Predictive Modelling ◽

Transmission Dynamics ◽

Background Rate ◽

Vaccination Rates ◽

Modelling Framework ◽

R Shiny ◽

Order Of Magnitude ◽

Main Components

BACKGROUND Modelling COVID-19 transmission at live events and public gatherings is essential to control the probability of subsequent outbreaks and communicate to participants their personalised risk. Yet, despite the fast-growing body of literature on COVID transmission dynamics, current risk models either neglect contextual information on vaccination rates or disease prevalence or do not attempt to quantitatively model transmission. OBJECTIVE This paper attempts to bridge this gap by providing informative risk metrics for live public events, along with a measure of their uncertainty. METHODS Building upon existing models, our approach ties together three main components: (a) reliable modelling of the number of infectious cases at the time of the event, (b) evaluation of the efficiency of pre-event screening, and (c) modelling of the event’s transmission dynamics and their uncertainty along using Monte Carlo simulations. RESULTS We illustrate the application of our pipeline for a concert at the Royal Albert Hall and highlight the risk’s dependency on factors such as prevalence, mask wearing, or event duration. We demonstrate how this event held on three different dates (August 3rd 2020, January 18th 2021, and March 8th 2021) would likely lead to transmission events only slightly above background rates (0.5 vs 0.2, 6.7 vs 3.5, and 5.4 vs 2.5, respectively. However, the 97.5 percentile of the prediction interval for the infections would likely be substantially higher than the background rate (6.8 vs 2, 89 vs 8, and 71 vs 7), further demonstrating that sole reliance on vaccination and antigen testing to gain entry would likely significantly underestimate the tail risk of the event. CONCLUSIONS Despite the unknowns surrounding COVID-19 transmission, our estimation pipeline opens the discussion on contextualized risk assessment by combining the best tools at hand to assess the order of magnitude of the risk. Our model can be applied to any future event, and is presented in a user-friendly R Shiny interface.

Download Full-text

Predicting ecological regime shift under climate change: New modelling techniques and potential of molecular-based approaches

Current Zoology ◽

10.1093/czoolo/59.3.403 ◽

2013 ◽

Vol 59 (3) ◽

pp. 403-417 ◽

Cited By ~ 5

Author(s):

Richard Stafford ◽

V. Anne Smith ◽

Dirk Husmeier ◽

Thomas Grima ◽

Barbara-ann Guinn

Keyword(s):

Climate Change ◽

Bayesian Network ◽

Species Interactions ◽

Rocky Shore ◽

Regime Shift ◽

Predictive Modelling ◽

Kelp Forests ◽

Stable Regime ◽

Ecological Regime ◽

Stable Community

Abstract Ecological regime shift is the rapid transition from one stable community structure to another, often ecologically inferior, stable community. Such regime shifts are especially common in shallow marine communities, such as the transition of kelp forests to algal turfs that harbour far lower biodiversity. Stable regimes in communities are a result of balanced interactions between species, and predicting new regimes therefore requires an evaluation of new species interactions, as well as the resilience of the ‘stable’ position. While computational optimisation techniques can predict new potential regimes, predicting the most likely community state of the various options produced is currently educated guess work. In this study we integrate a stable regime optimisation approach with a Bayesian network used to infer prior knowledge of the likely stress of climate change (or, in practice, any other disturbance) on each component species of a representative rocky shore community model. Combining the results, by calculating the product of the match between resilient computational predictions and the posterior probabilities of the Bayesian network, gives a refined set of model predictors, and demonstrates the use of the process in determining community changes, as might occur through processes such as climate change. To inform Bayesian priors, we conduct a review of molecular approaches applied to the analysis of the transcriptome of rocky shore organisms, and show how such an approach could be linked to meas-ureable stress variables in the field. Hence species-specific microarrays could be designed as biomarkers of in situ stress, and used to inform predictive modelling approaches such as those described here.

Download Full-text