Search

Large, varied, and time-evolving data sources can be observed across many domains and present a unique challenge for classification problems, in which traditional machine learning approaches must be adapted to accommodate for the complex nature of such data. Across most domains, there is also a need for machine learning models that are both well-performing and interpretable, to help provide explanations of a model's decisions that stakeholders can trust and take appropriate actions with. In the medical domain, complex Electronic Health Record (EHR) data consists of longitudinal records of patient histories spanning structured and unstructured data types. Exploiting such complex medical data is vital as a means of gaining useful medical insights and predictions, and where establishing stakeholder trust through useful explanations is critical. This thesis has focused on producing state-of-the-art classification methods for exploiting the heterogeneity and temporality of complex data; secondly, on developing novel interpretability methods to aid in the understanding of model predictions from such complex data; and finally on ensuring the medical applicability of the developed methods and other novel methods particularly for the medical problem of adverse drug event (ADE) prediction. In the first part of this thesis, several state-of-the-art classification frameworks for exploiting complex medical data are outlined, with their utility demonstrated through comparative empirical evaluations to competing framework approaches. In the second part of this thesis, novel interpretability methods are developed and demonstrated for their applicability across domains. In the third part of this thesis, the applicability of interpretability and explanability methods for complex medical data are investigated, refined, and assessed for validity in connection to the use-case of ADE prediction. Main contributions of this thesis include: two novel classification frameworks, including SM

Attention mechanisms form the basis of providing temporal explanations for a variety of state-of-the-art recurrent neural network (RNN) based architectures. However, evidence is lacking that attention mechanisms are capable of providing sufficiently valid medical explanations. In this study we focus on the quality of temporal explanations for the medical problem of adverse drug event (ADE) prediction by comparing explanations globally and locally provided by an attention-based RNN architecture against those provided by more a more basic RNN using the post-hoc SHAP framework, a popular alternative option which adheres to several desirable explainability properties. The validity of this comparison is supported by medical expert knowledge gathered for the purpose of this study. This investigation has uncovered that these explanation methods both possess appropriateness for ADE explanations and may be used complementarily, due to SHAP providing more clinically appropriate global explanations and attention mechanisms capturing more clinically appropriate local explanations. Additional feedback from medical experts reveal that SHAP may be more applicable to real-time clinical encounters, in which efficiency must be prioritised, over attention explanations which possess properties more appropriate for offline analyses.

Adverse drug events are pervasive and costly medical conditions, in which novel research approaches are needed to investigate the nature of such events further and ultimately achieve early detection and prevention. In this paper, we seek to characterize patients who experience an adverse drug event, represented as a case group, by contrasting them to similar control group patients who do not experience such an event. To achieve this goal, we utilize an extensive electronic patient record database and apply a combination of frequent arrangement mining and disproportionality analysis. Our results have identified how several adverse drug events are characterized in regards to frequent disproportional arrangements, where we highlight how such arrangements can provide additional temporal-based information compared to similar approaches.

A variety of deep learning architectures have been developed for the goal of predictive modelling in regards to detecting health diagnoses in medical records. Several models have placed strong emphases on temporal attention mechanisms and decay factors as a means to include highly temporally relevant information regarding the recency of medical event occurrence while facilitating medical code-level interpretability. In this study we utilise such models with a novel Electronic Patient Record (EPR) data set consisting of both diagnoses and medication data for the purpose of Adverse Drug Event (ADE) prediction. As such, a main contribution of this work is an empirical evaluation of two state-of-the-art deep learning architectures in terms of objective performance metrics for ADE prediction. We also assess the importance of attention mechanisms in regards to their usefulness for medical code-level interpretability, which may facilitate novel insights pertaining to the nature of ADE occurrence within the health care domain.

Cyrptocurrency price prediction has recently become an alluring topic, attracting massive media and investor interest. Traditional models, such as Autoregressive Integrated Moving Average models (ARIMA) and models with more modern popularity, such as Recurrent Neural Networks (RNN’s) can be considered candidates for such financial prediction problems, with RNN’s being capable of utilizing various endogenous and exogenous input sources. This study compares the model performance of ARIMA to that of a seq2seq recurrent deep multi-layer neural network (seq2seq) utilizing a varied selection of inputs types. The results demonstrate superior performance of seq2seq over ARIMA, for models generated throughout most of bitcoin price history, with additional data sources leading to better performance during less volatile price periods.

Over the last decades the healthcare domain has seen a tremendous increase and interest in methods for making inference about patient care using large quantities of medical data. Such data is often stored in electronic health records and administrative health registries. As these data sources have grown increasingly complex, with millions of patients represented by thousands of attributes, static or time evolving, finding relevant and accurate patterns that can be used for predictive or descriptive modelling is impractical for human experts. In this paper, we concentrate our review on Swedish Administrative Health Registries (AHRs) and Electronic Health Records (EHRs) and provide an overview of recent and ongoing work in the area with focus on adverse drug events (ADEs) and heart failure., QC 20180118