Publicação

Modeling the outbreak and spread of infectious diseases using a bayesian machine learning approach

Ver documento

Detalhes bibliográficos
Resumo:The modeling of infectious diseases and their predictions on space and time is very important as it helps in devising the policies for preventive measures. These predictions should be generated from a probabilistic model to provide the uncertainties and thus the confidence. The phenomenon of spread of infectious diseases is so complex that there are lots of uncertainties in the data and in the process itself. Machine learning methods like neural networks are useful in modeling this complex problem, however, these approaches lack handling of uncertainties. Similarly, it is seen in literature that a combined approach of neural networks and Bayesian inferences have not been explored much. Thus to fill these gaps this thesis aims to develop a combined model containing neural network method and Bayesian inference for modeling and predicting the number of cases of infectious diseases in areal units such as municipalities or health-zones. To introduce the impact of human movement on the spread of infectious disease, the movement data has been used combined with the daily infection data to form a spatial factor and used as a covariate in this study. In addition to this, the spatial correlation due to spatial neighborhood as well as the mobility is taken into account in the model along with the temporal dependencies. The model was evaluated on the COVID-19 dataset for 245 healthzones of the autonomous community of Castilla-Leon, Spain. The results show that the model is generally able to predict the number of cases of infectious diseases with good accuracy. Similarly, the mobility factor was also found to have an influence on the model. However, the flexibility of the model still needs to be evaluated by applying the model to different scenarios.
Autores principais:Niraula, Poshan
Assunto:Bayesian Inference Human movement Infectious diseases Neural networks SDG 3 - Good health and well-being
Ano:2021
País:Portugal
Tipo de documento:dissertação de mestrado
Tipo de acesso:acesso aberto
Instituição associada:Universidade Nova de Lisboa
Idioma:inglês
Origem:Repositório Institucional da UNL
Descrição
Resumo:The modeling of infectious diseases and their predictions on space and time is very important as it helps in devising the policies for preventive measures. These predictions should be generated from a probabilistic model to provide the uncertainties and thus the confidence. The phenomenon of spread of infectious diseases is so complex that there are lots of uncertainties in the data and in the process itself. Machine learning methods like neural networks are useful in modeling this complex problem, however, these approaches lack handling of uncertainties. Similarly, it is seen in literature that a combined approach of neural networks and Bayesian inferences have not been explored much. Thus to fill these gaps this thesis aims to develop a combined model containing neural network method and Bayesian inference for modeling and predicting the number of cases of infectious diseases in areal units such as municipalities or health-zones. To introduce the impact of human movement on the spread of infectious disease, the movement data has been used combined with the daily infection data to form a spatial factor and used as a covariate in this study. In addition to this, the spatial correlation due to spatial neighborhood as well as the mobility is taken into account in the model along with the temporal dependencies. The model was evaluated on the COVID-19 dataset for 245 healthzones of the autonomous community of Castilla-Leon, Spain. The results show that the model is generally able to predict the number of cases of infectious diseases with good accuracy. Similarly, the mobility factor was also found to have an influence on the model. However, the flexibility of the model still needs to be evaluated by applying the model to different scenarios.