This thesis introduces the first study of instance segmentation applied to gas flares in single beam echo sounder data. We develop a comprehensive dataset consisting of 1,414 images, featuring 5,142 segmented objects identified as gas flare. A key contribution is the adaptation of the Brier score specifically for instance segmentation. Further, we show how to adapt the Weighted Box Fusion (WBF) ...

Type 1 diabetes (T1D) is a chronic autoimmune disease that leads to insulin deficiency. Consequently, the disease will lead to a poor blood glucose (BG) regulation, and in situations of high energy expenditure like exercise, a dysfunctional regulation can cause severe damage or death [5] [14]. Diabetes is responsible for one death every five seconds and financially drains approximately 11% of the ...

This thesis is concerned with classification and regression using deep learning applied to fish otolith images. Otoliths (earstones) are calcified structures in the inner ear of vertebrates, and are used, for instance, in fish stock assessment and fish age determination. We use convolutional neural networks – a class of deep learning models - on two specific problems: discrimination between Northeast ...

Covid-19 has had a significant impact on daily life since the initial outbreak of the global pandemic in late 2019. Countries have been affected to varying degrees, depending on government actions and country characteristics such as infrastructure and demographics. Using Norway and Germany as a case study, this thesis aims to determine which factors influence the risk of infection in each country, ...

Trans-dimensional Bayesian inference for multi-layer perceptron architectures of varying size by reversible jump Markov chain Monte Carlo is developed and examined for its theoretical and practical merits and considerations. The algorithm features the No-U-Turn Sampler and Hamiltonian Monte Carlo for within-dimension moves, and makes use of a delayed-rejection sampler while exploring a variety of ...

The goal of this thesis is analyse the incidence rate of the first ever myocardial infarction (MI) and the survival time after the MI. The data used for this purpose is from the Troms\o{} study surveys collected in the period from 1974 to 2008. This thesis provides a general introduction to latent Gaussian models and the methodology of integrated nested Laplace approximation. Specifically, the data ...

This thesis aims to apply the Dirichlet process mixture model to the cluster kernel framework. The probabilistic cluster kernel is extended with a Bayesian nonparametric model to avoid critical parameters within the model. The Dirichlet process cluster kernel demonstrate advantages compared to the probabilistic cluster kernel in both classification and clustering. Additionally, the two dimensional ...

In this thesis we challenge the existence of weak efficiency in the Norwegian Stock Marked, by analysing time-lagged cross-correlations between log-return series from 811 stocks listed on the Oslo stock exchange and by creating prediction strategies based on the discovered patterns. We limit the strategies to predicting the direction of the movements of the time series only, i.e. either generating ...

In this thesis we generalize the Hansen and Seo test in the R package tsDyn, which tests a linear cointegration model against a two-regime threshold cointegration model, to the case of three regimes in the alternative hypothesis. As the Lagrange Multiplier test statistic used in the Hansen and Seo test in tsDyn is different from the LM statistic described in Hansen and Seo (2002), we generalize ...

All points on the surface of the Earth are moving. To define the velocity of a given point, we can place a GPS receiver there and measure the coordinates every day. After collecting enough data, we can generate a time series of three coordinates, North, East and Height directions. The most used technique to determine such displacements, is the linear model. The main objective of this thesis ...