Probabilistic graphical modeling (PGM) provides a framework for formulating an interpretable generative process of data and expressing uncertainty about unknowns. This makes PGM very useful for understanding phenomena underlying data and for decision making. PGM has seen great success in domains where interpretable inferences are key, e.g. marketing, medicine, neuroscience, and social science. However, PGM tends to lack flexibility, which has hindered its use when it comes to modeling large scale high-dimensional complex data and performing tasks that require flexibility (e.g. in vision and language applications.) Deep learning (DL) is another framework for modeling and learning from data that has seen great empirical success in recent years. DL is very powerful and offers great flexibility, but it lacks the interpretability and calibration of PGM. This thesis develops deep probabilistic graphical modeling (DPGM). DPGM consists in leveraging DL to make PGM more flexible. DPGM brings about new methods for learning from data that exhibit the advantages of both PGM and DL. We use DL within PGM to build flexible models endowed with an interpretable latent structure. One family of models we develop extends exponential family principal component analysis (EF-PCA) using neural networks to improve predictive performance while enforcing the interpretability of the latent factors. Another model class we introduce enables accounting for long-term dependencies when modeling sequential data, which is a challenge when using purely DL or PGM approaches. This model class for sequential data was successfully applied to language modeling, unsupervised document representation learning for sentiment analysis, conversation modeling, and patient representation learning for hospital readmission prediction. Finally, DPGM successfully solves several outstanding problems of probabilistic topic models. Leveraging DL within PGM also brings about new algorithms for learning with complex data. For example, we develop entropy-regularized adversarial learning, a learning paradigm that deviates from the traditional maximum likelihood approach used in PGM. From the DL perspective, entropy-regularized adversarial learning provides a solution to the long-standing mode collapse problem of generative adversarial networks.

Structural equation modelling (SEM) is carried out with the aim of testing hypotheses on the model of the researcher in a quantitative way, using the sampled data. Although SEM has developed in many aspects over the past few decades, there are still numerous advances which can make SEM an even more powerful technique. We propose representing the nal theoretical SEM by a Bayesian Network (BN), which we would like to call a Probabilistic Structural Equation Model (PSEM). With the PSEM, we can take things a step further and conduct inference by explicitly entering evidence into the network and performing di erent types of inferences. Because the direction of the inference is not an issue, various scenarios can be simulated using the BN. The augmentation of SEM with BN provides signi cant contributions to the eld. Firstly, structural learning can mine data for additional causal information which is not necessarily clear when hypothesising causality from theory. Secondly, the inference ability of the BN provides not only insight as mentioned before, but acts as an interactive tool as the `what-if' analysis is dynamic. / Mini Dissertation (MCom)--University of Pretoria, 2018. / Statistics / MCom / Unrestricted

In producing a user interface design to visualize search results for a digital library called Envision [Nowell, France, Hix, Heath, &amp; Fox, 1996] [Fox, Hix, Nowell, et al., 1993] [Nowell &amp; Hix, 1993], we found that choosing graphical devices and document attributes to be encoded with each graphical device is a surprisingly difficult task. By <i>graphical devices</i> we mean those visual display elements (e.g., color, shape, size, position, etc.) used to convey encoded, semantic information. Research in the areas of psychophysics of visual search and identification tasks, graphical perception, and graphical language development provides scientific guidance for design and evaluation of graphical encodings which might otherwise be reduced to opinion and personal taste. However, literature offers inconclusive and often conflicting viewpoints, suggesting a need for further research. The goal of this research was to determine empirically the effectiveness of graphical devices for encoding nominal and quantitative information in complex visualization displays. Using the Envision Graphic View, we conducted a within-subjects empirical investigation of the effectiveness of three graphical devices - <i>icon color, icon shape,</i> and <i>icon size</i> - in communicating nominal (document type) and quantitative (document relevance) data. Our study provides empirical evidence regarding the relative effectiveness of icon color, shape, and size for conveying both nominal and quantitative data. While our studies consistently rank color as most effective, the rankings differ for shape and size. For nominal data, icon shape ranks ahead of icon size by all measures except time for task completion, which places shape behind size. For quantitative data, we found, by all measures, that encodings with icon shape are more effective than with icon size. We conclude that the <i>nature of tasks</i> performed and the relative <i>importance of measures of effectiveness</i> are more significant than the type of data represented for designers choosing among rankings. / Ph. D.

In the mobile era, data capturing individuals’ locations have become unprecedentedly available. Data from Location-Based Social Networks is one example of large-scale user-location data. Such data provide a valuable source for understanding patterns governing human mobility, and thus enable a wide range of research. However, mining and utilizing raw user-location data is a challenging task. This is mainly due to the sparsity of data (at the user level), the imbalance of data with power-law users and locations check-ins degree (at the global level), and more importantly the lack of a uniform low-dimensional feature space describing users. Three latent feature models are proposed in this dissertation. Each proposed model takes as an input a collection of user-location check-ins, and outputs a new representation space for users and locations respectively. To avoid invading users privacy, the proposed models are designed to learn from anonymized location data where only IDs - not geophysical positioning or category - of locations are utilized. To enrich the inferred mobility patterns, the proposed models incorporate metadata, often associated with user-location data, into the inference process. In this dissertation, two types of metadata are utilized to enrich the inferred patterns, timestamps and social ties. Time adds context to the inferred patterns, while social ties amplifies incomplete user-location check-ins. The first proposed model incorporates timestamps by learning from collections of users’ locations sharing the same discretized time. The second proposed model also incorporates time into the learning model, yet takes a further step by considering time at different scales (hour of a day, day of a week, month, and so on). This change in modeling time allows for capturing meaningful patterns over different times scales. The last proposed model incorporates social ties into the learning process to compensate for inactive users who contribute a large volume of incomplete user-location check-ins. To assess the quality of the new representation spaces for each model, evaluation is done using an external application, social link prediction, in addition to case studies and analysis of inferred patterns. Each proposed model is compared to baseline models, where results show significant improvements.

Providing consumers with usable nutrition information requires an effective labeling format. The objective of this study, which was conducted in a supermarket setting, was to determine whether consumers could, without previous instruction, make equally effective nutrition decisions using a graphic format based on nutrient density as when using the current labeling format. For comparison with other studies, a demographic, nutrition knowledge and nutrition labeling data base was collected. The questionnaire completed by each participating consumer included items regarding demographic data and shopping preferences, and questions evaluating nutrition knowledge for comparison as a data base with other studies. It also appraised the ability of the shopper to utilize two nutrition labeling formats in making nutrition decisions. Another questionnaire, completed by a researcher, assessed race, body type and build, and time taken by each participant to complete the nutrition decision questions. Six supermarkets were selected from one large Utah chain as sites for the survey. The nutrient density format produced the greatest percentage of correct responses. The difference was particularly evident when the data were analyzed for overall total correct responses. Participants who were high school graduates or had family incomes between $4,000 - $7,999 made more correct responses when utilizing the nutrient density format than when confronted with the other format. The nutrient density presentation also took less time for participants to complete. The graphical nutrient density format evaluated in the study is more effective than the current labeling format in assisting consumers to make valid nutritional decisions.

Previous research has suggested that misinterpretations of hurricane track forecasts can lead to errors in estimation of perceived risk. One factor that can be used to understand these errors in judgment of risk perception is called optimistic bias, in which an individual perceives that compared to another person they are at less risk. Thus, the purpose of this study was to examine how risk perceptions of hurricane track forecasts are influenced by the optimistic bias and changes in the forecasts. Students from three coastal universities took a survey regarding hurricane risk from two different track scenarios of a hypothetical hurricane approaching their university. Results indicated that optimism and perceptions of hurricane tracks were not correlated. Regardless of changes in forecast tracks, students perceived the same level of risk by the final forecast. This research has important social implications because hurricane track forecasts are part of the hurricane decision-making process.

An influence diagram is a widely-used graphical model for representing and solving problems of sequential decision making under imperfect information. A closely-related model for the same class of problems is a partially observable Markov decision process (POMDP). This dissertation leverages the relationship between these two models to develop improved algorithms for solving influence diagrams. The primary contribution is to generalize two classic dynamic programming algorithms for solving influence diagrams, Arc Reversal and Variable Elimination, by integrating them with a dynamic programming technique originally developed for solving POMDPs. This generalization relaxes constraints on the ordering of the steps of these algorithms in a way that dramatically improves scalability, especially in solving complex, multi-stage decision problems. A secondary contribution is the adoption of a more compact and intuitive representation of the solution of an influence diagram, called a strategy. Instead of representing a strategy as a table or as a tree, a strategy is represented as an acyclic graph, which can be exponentially more compact, making the strategy easier to interpret and understand.

Achilles tendonbrott (Achilles Tendon Rupture, ATR) är en av de typiska mjukvävnadsskadorna. Rehabilitering efter sådana muskuloskeletala skador förblir en långvarig process med ett mycket variet resultat. Att kunna förutsäga rehabiliteringsresultat exakt är avgörande för beslutsfattande stöduppdrag. I detta arbete designar vi en probabilistisk modell för att förutse rehabiliteringsresultat för ATR med hjälp av en klinisk kohort med många saknade poster. Vår modell är tränad från början till slutet för att samtidigt förutsäga de saknade inmatningarna och rehabiliteringsresultat. Vi utvärderar vår modell och jämför med flera baslinjer, inklusive flerstegsmetoder. Experimentella resultat visar överlägsenheten hos vår modell över dessa flerstadiga tillvägagångssätt med olika dataimuleringsmetoder för ATR rehabiliterings utfalls prognos. / Achilles Tendon Rupture (ATR) is one of the typical soft tissue injuries. Rehabilitation after such musculoskeletal injuries remains a prolonged process with a very variable outcome. Being able to predict the rehabilitation outcome accurately is crucial for treatment decision support. In this work, we design a probabilistic model to predict the rehabilitation outcome for ATR using a clinical cohort with numerous missing entries. Our model is trained end-to-end in order to simultaneously predict the missing entries and the rehabilitation outcome. We evaluate our model and compare with multiple baselines, including multi-stage methods. Experimental results demonstrate the superiority of our model over these baseline multi-stage approaches with various data imputation methods for ATR rehabilitation outcome prediction.

The statistical modeling of spatial extreme events, augmented by graphical models, provides a comprehensive framework for the development of techniques and models to describe natural phenomena in a variety of environmental, geoscience, and climate science applications. In a changing climate, the impact of natural hazards, such as wildfires, is believed to have evolved in frequency, size, and spatial extent, although regional responses may vary. The aforementioned impacts are of great significance due to their association with air pollution, irreversible harm to the environment and atmosphere, and the fact that they put human lives at risk. The prediction of wildfires holds significant importance within the realm of wildfire management due to its influence on the allocation of resources, the mitigation of detrimental consequences, and the subsequent recovery endeavors. Therefore, the development of robust statistical methodologies that can accurately forecast extreme wildfire occurrences across spatial and temporal dimensions is of great significance. In this thesis, we develop new spatial statistical models, combined with popular machine learning techniques, as well as novel extreme-value methods to enhance the prediction of wildfire risk using graphical models. First, in order to jointly efficiently model high-dimensional wildfire counts and burnt areas over the whole continguous United States, we propose a four-stage zero-inflated bivariate spatiotemporal model combining low-rank spatial models and random forests. Second, to model high values of the McArthur Forest Fire Danger Index over Australia, we develop a novel spatial extreme-value model based on mixtures of tree-based multivariate Pareto distributions. Our new methodology combines theoretically justified spatial extreme models with a computationally convenient graphical model framework to spatial problems in high dimensions efficiently. Third, we exploit recent advancements in deep learning and build a parametric regression model using graphic convolutional neural networks and the extended Generalized Pareto distribution, allow us to jointly model moderate and extreme wildfires observed on irregular spatial grid. We work with a novel dataset of Australian wildfires from 1999 to 2019, and analyse monthly spread over areas correspond to Statistical Area Level 1 regions. We highlight the efficacy of our newly proposed model and perform risk assessment for Australia and dense communities.

This research focuses on the integration of Eye Movement Desensitization and Reprocessing therapy (EMDR) and Virtual Reality (VR). EMDR is a therapeutic approach using bilateral stimulation to process distressing memories, emotions, and experiences. It is widely employed for conditions like PTSD, anxiety, and depression. The study involves the development of a proof-of-concept virtual reality tool tailored for EMDR therapy sessions. The tool comprises a bilaterally moving sphere within the user's VR environment, controllable by the therapist through a graphical interface on a computer. The therapist can dynamically adjust the sphere's color, speed, and sound to enhance the therapeutic process. The study's findings affirm the feasibility of creating a VR tool that supports therapists in conducting effective EMDR therapy sessions.