Date of Award

12-19-2023

Publication Type

Thesis

Degree Name

M.Sc.

Department

Computer Science

Keywords

COVID-19 epidemic prediction;Deep Learning;Machine Learning;Multi Variate Multi Timeseries Dataset;Multi Variate Time Series;Time Series

Supervisor

Ziad Kobti

Rights

info:eu-repo/semantics/openAccess

Abstract

This thesis introduces an innovative framework aimed at addressing the complexities of predicting outcomes in multivariate multi time series datasets in regression analysis. By applying this framework to a novel COVID-19 dataset, it enhances predictive analytics by providing accurate forecasts for epidemic trends at regional or provincial levels, going beyond national-level analysis. The framework incorporates advanced data preprocessing, feature selection, engineering, encoding, and model architecture, effectively capturing intricate variable interactions and temporal dependencies. This makes it a powerful tool for tackling multivariate multi time series regression challenges, offering valuable insights for informed decision-making. Predicting outcomes in such datasets is challenging due to variable interconnections and temporal dynamics. The framework presented in the thesis adeptly models dependencies and latent patterns while considering real-world uncertainties. It demonstrates its practical value in localized epidemic trend forecasting, where deep data understanding is crucial for effective decision-making. Extensive experimentation shows that the framework outperforms traditional regression models and time series models in terms of various performance metrics, such as $R^2$, MAE, MaxAE, and RMSE. A novel model, DeepAREstimator, is introduced to balance performance and training time, offering a maintainable and scalable solution for real-world applications. The findings contribute to advancing predictive analytics, and providing essential insights for decision-making, particularly in localized epidemic trend forecasting.

Download

Share

COinS