Advanced information technologies promise a massive influx of individual-specific medical data. These rich sources offer great potential for an increased understanding of disease mechanisms and for providing evidence-based and personalized clinical decision support. However, the size, complexity, and biases of the data pose new challenges, which make it difficult to transform the data to useful and actionable knowledge using conventional statistical analysis. The so-called “Big Data” era has created an emerging and urgent need for scalable, computer-based data mining methods that can turn data into useful, personalized decision support knowledge in a flexible, cost-effective, and productive way. The goal of my Ph.D. research is to address some key challenges in current clinical deci-sion support, including (1) the lack of a flexible, evidence-based, and personalized data mining tool, (2) the need for interactive interfaces and visualization to deliver the decision support knowledge in an accurate and effective way, (3) the ability to generate temporal rules based on patient-centric chronological events, and (4) the need for quantitative and progressive clinical predictions to investigate the causality of targeted clinical outcomes. The problem statement of this dissertation is that the size, complexity, and biases of the current clinical data make it very difficult for current informatics technologies to extract individual-specific knowledge for clinical decision support. This dissertation addresses these challenges with four overall specific aims: Evidence-Based and Personalized Decision Support: To develop clinical decision support systems that can generate evidence-based rules based on personalized clinical conditions. The systems should also show flexibility by using data from different clinical settings. Interactive Knowledge Delivery: To develop an interactive graphical user interface that expedites the delivery of discovered decision support knowledge and to propose a new visualiza-tion technique to improve the accuracy and efficiency of knowledge search. Temporal Knowledge Discovery: To improve conventional rule mining techniques for the discovery of relationships among temporal clinical events and to use case-based reasoning to evaluate the quality of discovered rules. Clinical Casual Analysis: To expand temporal rules with casual and time-after-cause analyses to provide progressive clinical prognostications without prediction time constraints. The research of this dissertation was conducted with frequent collaboration with Children’s Healthcare of Atlanta, Emory Hospital, and Georgia Institute of Technology. It resulted in the development and adoption of concrete application deliverables in different medical settings, including: the neuroARM system in pediatric neuropsychology, the PHARM system in predictive health, and the icuARM, icuARM-II, and icuARM-KM systems in intensive care. The case studies for the evaluation of these systems and the discovered knowledge demonstrate the scope of this research and its potential for future evidence-based and personalized clinical decision support.

Thesis (PhD)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: The policing and monitoring of South Africa's coastline and economic exclusion zone is made di cult not only by the size of the area of interest, but also by the limited resources available for maritime detection and policing. As a consequence, illegal activities, such as smuggling, poaching and illegal border crossings, are often conducted with impunity. Conventional approaches to monitoring coastal areas, such as the use of patrol boats, port inspections and aircraft surveillance, may be augmented by advances in technology that are steadily contributing vast amounts of data related to maritime activity. For example, various South African agencies collect auto- matic identi cation system and vessel monitoring system transmissions, and gather additional kinematic data of maritime vessels through a number of strategically placed coastal radars. A command and control centre for actively monitoring these data (outside of the intelligence community) was established by the South African Navy in 2014. Such centres provide surveillance operators with a real-time picture of a maritime region of interest from which they can identify relevant facts of interest through a reliance on experience and domain knowledge. The e ectiveness of this process may, however, be undermined by the vast quantities of data typically under consideration, by the di culty of identifying long-term trends in vessel kinematic behaviour and by the possibility of operator fatigue brought on by the relatively low incidence levels of activities of interest. E ective decision support tools may play a valuable role in this context by the automatic processing of these vast collections of data, by the identi cation of concepts of interest and by the prediction of future occurrences of interest. It is, however, essential that such tools should be exible enough to adapt to changes in typical vessel behaviour over time and that they should be capable of integrating new trends and new types of behaviours. Various approaches to maritime surveillance are investigated in this dissertation from the perspectives of threat detection and anomaly identi cation, with particular emphasis on a systems approach to decision support. A decision support system framework that utilises rule-based and data-driven mechanisms is proposed as a means to separate the interesting from the uninteresting and to provide early warnings of potentially threatening maritime vessel behaviour to operators. This system framework is primarily concerned with kinematic data and is restricted to the identi cation of certain types of activities. Successful classi cation and, ultimately, timely prediction of potentially threatening behaviour would allow for e ective policing by providing early warning to relevant entities, thus potentially leading to more e ective use of available policing resources. / AFRIKAANSE OPSOMMING: Die patrollering en monitering van die Suid-Afrikaanse kusgebied en gepaardgaande ekonomiese eksklusiewe zone word bemoeilik deur die grootte van die tersprake area en die beperkte hulpbronne wat vir patrollie-doeleindes aangewend kan word. Gevolglik gaan onwettige aktiwiteite, soos smokkelary, stroping en onwettige immigrasie dikwels ongestraf. Konvensionele benaderings tot die monitering van kusgebiede, soos die aanwending van patrolliebote, die uitvoer van hawe-inspeksies en gere elde lugpatrollies, kan aangevul word deur tegnologiese vooruitgang wat voortdurend tot groot hoeveelhede data oor maritieme aktiwiteit bydra. Verskeie Suid- Afrikaanse agentskappe ontvang byvoorbeeld outomatiese identi kasiestelsel en vaartuigmoni- teringstelsel uitsendings, en samel ook addisionele kinematiese data oor maritieme vaartuie deur middel van strategies-geplaasde kusradars in. 'n Bevel-en-beheersentrum wat hierdie inligting (buite die intelligensiegemeenskap) aktief ontleed, is in 2014 deur die Suid-Afrikaanse Vloot tot stand gebring. Sulke sentra verskaf 'n intydse blik oor die maritieme gebied onder beskouing aan operateurs wat dan, gebaseer op hulle ervaring en omgewingskennis, relevante inligting oor vaartuie kan a ei. Die doeltre ende uitvoering van hierdie proses kan egter ondermyn word deur die tipiese groot hoeveelhede data, die moeilikheidsgraad van die identi kasie van langtermyn tendense in die kinematiese gedrag van vaartuie om die kus en die moontlikheid van operateur-uitputting as gevolg van lang periodes van relatiewe oninteressante vaartuiggedrag. Doeltre ende besluitsteunhulpmiddels kan 'n waardevolle bydrae in hierdie konteks maak deur die ge-outomatiseerde prosessering van hierdie groot hoeveelhede data, die identi kasie van interessante vaartuiggedrag en die voorspelling van toekomstige relevante insidente. Dit is egter noodsaaklik dat sulke hulpmiddels buigsaam genoeg moet wees om te kan aanpas by veranderings in tipiese maritieme aktiwiteit oor tyd en dat nuwe tendense en tipes aktiwiteite geakkommodeer kan word. Verskeie benaderings tot maritieme oorsig word in hierdie proefskrif vanuit die perspektiewe van die bespeuring van bedreigings en die opsporing van vreemde verskynsels ondersoek, met 'n spesi eke fokus op 'n stelselbenadering tot besluitsteun. 'n Besluitsteun stelselraamwerk wat berus op re el-gebaseerde en data-aangedrewe meganismes word as 'n hulpmiddel voorgestel waarmee interessante maritieme gedrag van oninteressante gedrag onderskei kan word om sodoende 'n vroe e waarskuwing aan operateurs met betrekking tot moontlike bedreigende maritieme aktiwiteite te kan rig. Die werking van hierdie stelselraamwerk berus hoofsaaklik op die gebruik van kinematiese vaartuigdata en is beperk tot die naspeuring van sekere soorte bedreigende gedrag. Die suksesvolle klassi kasie en tydige voorspelling van potensi ele bedreigende maritieme gedrag behoort doeltre ende kusmonitering en verbeterde aanwending van die beperkte, gepaardgaande hulpbronne deur relevante kusagentskappe moontlik te maak.

published_or_final_version / Urban Planning and Environmental Management / Doctoral / Doctor of Philosophy

published_or_final_version / Business Administration / Master / Master of Business Administration

published_or_final_version / Business Administration / Master / Master of Business Administration

published_or_final_version / Business Administration / Master / Master of Business Administration

published_or_final_version / Industrial and Manufacturing Systems Engineering / Master / Master of Philosophy

The amount of information generated and maintained by information systems and their users leads to the increasingly important concern of information overload. Personalized systems have thus emerged to help provide more relevant information and services to the user. In particular, recommender systems appeared in the mid 1990's and have since then generated a growing interest in both industry and academia. Besides, context-aware systems have been developed to model, capture and interpret information about the user's situation, generally in dynamic and heterogeneous environments. Decision support systems like Business Intelligence (BI) platforms also face usability challenges as the amount of information available to knowledge workers grows. Remarkably, we observe that only a small part of personalization and recommendation techniques have been used in the context of data warehouses and analysis tools. Therefore, our work aims at exploring synergies of recommender systems and context-aware systems to develop personalization and recommendation scenarios suited in a BI environment. In response to this, we develop in our work an open and modular situation management platform using a graph-based situation model. Besides, dynamic aspects are crucial to deal with context data which is inherently time-dependent. We thus define two types of active components to enable dynamic maintenance of situation graphs, activation rules and operators. In response to events which can describe users' interactions, activation rules - defined using the event-condition-action framework - are evaluated thanks to queries on underlying graphs, to eventually trigger appropriate operators. These platform and framework allow us to develop and support various recommendation and personalization scenarios. Importantly, we design a re-usable personalized query expansion component, using semantics of multi-dimensional models and usage statistics from repositories of BI documents like reports or dashboards. This component is an important part of another experimentation we realized, Text-To-Query. This system dynamically generates multi-dimensional queries to illustrate a text and support the knowledge worker in the analysis or enrichment of documents she is manipulating. Besides, we also illustrate the integration and usage of our graph repository and situation management frameworks in an open and extensible federated search project, to provide background knowledge management and personalization.

Aided by advances in information technology, decision support systems (DSS) are widely used throughout organizations. These DSS are limited to support of specific structured and semi-structured management tasks for individual decision-makers and use primarily quantitative models. The next step in the evolution of DSS is to support complex, unstructured decision processes using qualitative, creativity enhancement models. The purpose of this research was to design, implement and evaluate an automated system to support complex, unstructured group decision processes. Idea generation and management in organization planning has been chosen as the domain for the system. A DSS architecture has been developed that includes a process management system component in addition to traditional data, dialogue and model management components. A group DSS and knowledge-based management system approach are central features of the system architecture. Software engineering methods were used to design, implement and evaluate the technical feasibility of the prototype system. Action research using participant and structured observation methods was used to study the (1) dynamics of the idea generation process during automated brainstorming, (2) influence of the technology on the idea generation process and (3) satisfaction of the planners with automated brainstorming for idea generation in a group setting. The findings of the research indicate that automated, networked idea generation can assist groups of planners in generating ideas during planning sessions. These ideas were accurately represented and stored and efficiently retrieved using a semantic inheritance network and frame knowledge management system implemented using a specially-designed knowledge representation language developed by the author. Over 100 planners from a variety of organizations used the system. Data indicated that computer brainstorming changed group dynamics, especially group interaction and participation. The anonymity provided by EBS neutralized social inhibitions and, in combination with the capability for parallel idea generation on the computer network, helped equalize participation. Minimal group interaction occurred. Planners using interactive computer brainstorming reported high levels of satisfaction with the process and outcome of the planning session.

Information systems have become an essential part of every business organization's production and management process. It is critical to an enterprise to integrate its organization and information systems. However, the lack of computer-supported tools for modeling organization and information systems has put their integration far beyond our reach. In this research, a metasystem approach that can integrate organization and information system modeling by means of group decision support systems (GDSS) and computer-aided software engineering (CASE) has been proposed. A prototype system, called MetaPlex, has been designed and implemented to demonstrate the feasibility of the proposed approach. The emphasis in design and implementation of MetaPlex has been on making the underlying knowledge representation expressive enough to meet modeling requirements and ensuring that the user interface is easy for managers and users to use. The use of a GDSS makes it possible to capture strategic assumptions and business objectives, as well as structures of an organization, from managers through face-to-face group meetings. The application of the metasystem concepts in generating GDSS tools makes the customization of a GDSS environment possible. Because of GDSS environment driven by a metasystem can be used to acquire information about a target system from multiple experts in a structured format that can be integrated with CASE tools, this approach provides a basis for a seamless integration of GDSS and CASE tools to support both organization and information system modeling.