All Projects

The accurate forecasting of healthcare demand is critical for ensuring optimal resource allocation, minimizing costs, and improving patient outcomes. This research aims to develop a robust demand forecasting model tailored to healthcare systems, using time-series modelling. By analyzing historical patient data, seasonal trends, demographic variables, and disease patterns, the model will predict future healthcare demands, … Read more

Bayesian Networks (BNs) have been widely used in decision science. BNs are trained using an available dataset where the probabilistic relationship between its components is defined. This project aims to study and further develop using BNs where the available data in one (or some) of the BN node(s) is censored.

A series probability distribution can be easily developed by dividing the terms of the Mac-Lauran expansion of a function f(x), by f(x). This provides a series with a sum of 1, which therefore can be assumed as a probability function. The above simple definition of a series distribution shows the flexibility of them, and therefore … Read more

Nowadays, Workforce Planning (WP) is a hot topic of interest if large state organizations. WP models are usually predictive (, econometric) time series models which aim to include various dynamic inputs to predict the availability of workforce. This analysis can be more extended by applying sensitivity analysis that assists the policy makers for better decision … Read more

GWR models have been widely used in the field of Spatial Statistics, where a global regression model seem to have not the required prediction precision and therefore the local data are decided to be used in GWR models. This project aims to study and further extend the theory and application of the GWR when the … Read more

This research aims to review/develop the analysis methodology/framework for modelling categorical (binary in the first instance) outputs in different timepoints using time-series data.

Inter-rater reliability or agreement can be thought of the consistency measure if a fixed number of raters assign numerical ratings to a number of cases/patients. This project aims to review the availsble inter-rater reliability measures and provide an expansion where required.

Vehicular ad hoc networks, deep-space interplanetary networks, sparse sensornetworks, underwater networks, and mobile social networks are wirelessnetworks that greatly benefit from DTN. Primarily, these networks are not connected, hencerouting techniques that rely on the end-to-end path, such as dynamic source routing (DSR)or ad hoc on-demand distance vector (AODV) cannot be utilized [10]. For successful datatransmission … Read more

Copulas have been widely used in various field of statistical analysis, such as Finance, where modelling of competing risks (random factors) is required. This project aims to study and further develop the application of Copulas in a different context (such as medical and epidemiological studies).

The Poisson Regression and Negative Binomial Regression models are the conventional statistical models for count data. This research aims to use Decision Trees (including some extended models such as Random Forest) in modelling of count data and comparing the performance with Poisson Regression and Negative Binomial Regression models.