Algorithms for repeated (or “online”) decision-making are predominantly studied under the assumption that feedback is either statistical (determined by fixed probability distributions) or adversarial (changing over time in a potentially worst-case manner). Both of these assumptions ignore a phenomenon commonly present in repeated interactions with other agents, in which the space of our possible future outcomes is shaped in a structured and potentially predictable manner by our history of prior decisions. In this thesis, we consider online decision problems where the feedback model is adaptive rather than purely statistical or adversarial. One such example is a repeated game played against an opponent who uses a learning algorithm of their own; here, we give a characterization of possible outcome spaces which unifies disparate equilibrium notions, and serves as a basis for designing new algorithms. We then consider the task of providing recommendations to an agent whose preferences adapt based on the recommendation history, where we explore algorithmic tradeoffs in terms of the structure of this adaptivity pattern. We conclude by offering a general framework and algorithmic toolkit for approaching adaptive problems of this form.

Since the 1990 the use of Electronic Patient Records (EPR) in health services has become increasingly prevalent world wide. EPR has become an important aspect of the continuous improvement of patient care. Transferring all patient records from paper based to electronic is now a priority for many health services. The research reported in this thesis is sponsored by Hamad Medical Corporation (HMC) to provide opportunity to explore the potential role for EPR in the Medical Records Department. The study has been designed to gain better understanding of the users perspectives with regard to the use of patient records. In order to analyse and understand the complex dynamic involved in the management and use of patient records, it was recognised that systems thinking offered an appropriate framework for this research. Soft System Methodology (SSM) was therefore applied to the analysis of the data and used to inform the development of a conceptual model. Using SSM in combination with the structured questionnaire survey and telephone semi-structured interview, triangulation of methods was achieved. Use of these generated rich data revealing for example the general dissatisfaction expressed with the existing manual patient records system, the lack of confidentiality, poor legibility, shortage of space and the frequent misfiling of records. The need to address these problems has informed the strategic plan for the development and implementation of EPR for HMC. The research has successfully addressed the stated aims and research questions and guided the formulation of proposals for improvements.

This quantitative research study investigated the impact of organisational context on the process and success of strategic IS planning (SISP) in post-implementation information systems in Libyan organisations. A set of direct and indirect relationships were investigated in the research model. The organisational context presented as a contingent situational variable mediated by SISP process and predicted by SISP success (the criterion variable). The causality of the relationship set was developed from the contingency theory of information systems and supported by fit models in strategic management research. The study deployed multivariate analysis represented in the structural equation modelling (SEM) to develop robust construct measurements and analyse data collected from executives responsible for information systems planning in both public and private Libyan organisations. Multi-dimensional multi-items constructs were used in the path analysis model after they were extensively validated. The path analysis model represented as mediation model, where hypothesise suggest that SISP context has an impact SISP success, through the influence of the SISP process. In the model, four dimensions of the SISP context construct were found to have a significant impact on SISP success directly and indirectly through the SISP process. Two of these dimensions are components of the leadership orientation construct, namely “Creative and Controlling” leadership. The other two dimensions are “Organisation centralisation structure and the Riskiness of organisation strategies”. The environmental uncertainty and planning resource constructs were found to have no impact on SISP success in Libyan organisations. Furthermore, this study validated six out of seven dimensions of SISP process construct measurement; only five exhibited acceptable fit level in the path analysis model and all were affected by the SISP context. However, just three out of five SISP process constructs had an impact on SISP success namely “Comprehensiveness, Focus and Intuition planning process”. Different SISP processes were associated with different levels of SISP success, “Intuition” was the most effective SISP process approach. The second most effective SISP process approach was the “Focus on innovation”, followed by “Limited comprehensiveness”. The SISP success measured by the fulfilment of key objectives that has three measurements constructs namely “Analysis, Alignment, and Cooperation”. The research suggest that under the effect of organisation context the most successful SISP produced by (CIO, CEO, or top executives) who rely less on personal judgment, focus more on innovation rather than control and limit their comprehensiveness of information systems planning process.

Bayesian probabilistic modeling provides a powerful framework for building flexible models to incorporate latent structures through likelihood model and prior. When we specify a model, we make certain assumptions about the underlying data-generating process with respect to these latent structures. For example, the latent Dirichlet allocation (LDA) model assumes that when generating a document, we first select a latent topic and then select a word that often appears in the selected topic. We can uncover the latent structures conditioned on the observed data via posterior inference. In this dissertation, we apply the tools of probabilistic latent variable models and try to understand complex real-world data about music semantics and user behavior. We first look into the problem of automatic music tagging -- inferring the semantic tags (e.g., "jazz'', "piano'', "happy'', etc.) from the audio features. We treat music tagging as a matrix completion problem and apply the Poisson matrix factorization model jointly on the vector-quantized audio features and a "bag-of-tags'' representation. This approach exploits the shared latent structure between semantic tags and acoustic codewords. We present experimental results on the Million Song Dataset for both annotation and retrieval tasks, illustrating the steady improvement in performance as more data is used. We then move to the intersection between music semantics and user behavior: music recommendation. The leading performance in music recommendation is achieved by collaborative filtering methods which exploit the similarity patterns in user's listening history. We address the fundamental cold-start problem of collaborative filtering: it cannot recommend new songs that no one has listened to. We train a neural network on semantic tagging information as a content model and use it as a prior in a collaborative filtering model. The proposed system is evaluated on the Million Song Dataset and shows comparably better result than the collaborative filtering approaches, in addition to the favorable performance in the cold-start case. Finally, we focus on general recommender systems. We examine two different types of data: implicit and explicit feedback, and introduce the notion of user exposure (whether or not a user is exposed to an item) as part of the data-generating process, which is latent for implicit data and observed for explicit data. For implicit data, we propose a probabilistic matrix factorization model and infer the user exposure from data. In the language of causal analysis (Imbens and Rubin, 2015), user exposure has close connection to the assignment mechanism. We leverage this connection more directly for explicit data and develop a causal inference approach to recommender systems. We demonstrate that causal inference for recommender systems leads to improved generalization to new data. Exact posterior inference is generally intractable for latent variables models. Throughout this thesis, we will design specific inference procedure to tractably analyze the large-scale data encountered under each scenario.

<p>Основни циљ истраживања представља повећање ефективности и продуктивности предузећа развојем модела за оцену успеха система за колаборацију према специфичним својствима индустрије, на основу евалуације различитих фактора платформи за колаборацију у реалним условима и њиховог утицаја на перформансе предузећа.<br />Утврђен је међузависни однос елемената модела система за колаборацију и утицај на перформансе предузећа.</p> / <p>Osnovni cilj istraživanja predstavlja povećanje efektivnosti i produktivnosti preduzeća razvojem modela za ocenu uspeha sistema za kolaboraciju prema specifičnim svojstvima industrije, na osnovu evaluacije različitih faktora platformi za kolaboraciju u realnim uslovima i njihovog uticaja na performanse preduzeća.<br />Utvrđen je međuzavisni odnos elemenata modela sistema za kolaboraciju i uticaj na performanse preduzeća.</p> / <p>The main objective of this study is to increase the effectiveness and productivity of companies by developing model for assessing the success of the collaboration system to the specific characteristics of the industry, based on the evaluation of various factors of platform for collaboration in real conditions and their impact on company performance.<br />The interdependent relationship between elements of the collaboration system and its impact on company&rsquo;s performances is determined.</p>

University of Macau / Faculty of Science and Technology / Department of Computer and Information Science

Thesis (M.Info.Tech.-Res.)--University of Wollongong, 2007. / Typescript. Includes bibliographical references.

Με αυτή την εργασία προτείνουμε και υποστηρίζουμε ένα νέο πρότυπο για τα συστήματα αρχείων νέας γενιάς. Σε αυτό το πρότυπο, η παραδοσιακή άποψη ενός αρχείου αντικαθίσταται από την έννοια της πληροφοριακής μονάδας (information unit) και η παραδοσιακή αντίληψη των ιεραρχικών συστημάτων αρχείων αντικαθίσταται από ένα συνεχώς εξελισσόμενο χώρο δυναμικά αλληλένδετων πληροφοριακών μονάδων. Ένα Οργανικό Σύστημα Αρχείων (OFS) ορίζεται ως ένα σύστημα το οποίο αναπτύσσεται φυσικά και δεν υπόκειται σε τεχνητούς κανόνες και προκαθορισμένους, στατικούς τρόπους προβολής προς τους χρήστες του. Στο επίκεντρο του OFS βρίσκονται νέες αφαιρέσεις που υποστηρίζουν ένα συνεχώς εξελισσόμενο σύνολο πληροφοριακών μονάδων, χαρακτηρισμών των χρηστών για αυτές, και σχέσεων που δημιουργούνται μεταξύ τους από την πρόσβαση των χρηστών σε αυτές. Οι αφαιρέσεις αυτές επιτρέπουν το ίδιο σύστημα και το περιεχόμενό του να είναι ορατό με διαφορετικό τρόπο από διαφορετικούς τύπους χρηστών, σύμφωνα με τις τρέχουσες πληροφοριακές τους ανάγκες. Το OFS είναι ανθρωποκεντρικό, καθώς απαιτείται ανθρώπινη συνεισφορά για το χαρακτηρισμό των πληροφοριακών μονάδων και για την ανακάλυψη και το σχολιασμό των μεταξύ τους σχέσεων. Δεδομένου αυτού, στην καρδιά του OFS βρίσκονται αλγόριθμοι για την αναζήτηση βάσει περιεχομένου στα αποθηκευμένα αρχεία. Στην εργασία αυτή εκθέτουμε τα αποτελέσματα της μέχρι τώρα έρευνάς μας, συμπεριλαμβανομένης μιας υλοποίησης σε επίπεδο πυρήνα του λειτουργικού συστήματος, των βασικών χαρακτηριστικών του OFS, καθώς και τις σχετικές μετρήσεις απόδοσης προς απόδειξη της βιωσιμότητας της προσέγγισής μας. Συζητάμε στη συνέχεια, τις προκλήσεις που παραμένουν και τον αντίκτυπο που μπορεί να έχει το OFS στις σχετικές προσπάθειες έρευνας και ανάπτυξης, επισημαίνοντας τη σχετική έρευνα από άλλους τομείς, όπως η Ανάκτηση Πληροφορίας, το Κοινωνικό Λογισμικό, οι Διεπαφές Χρηστών, και η Διαχείριση Δεδομένων. / We propose and advocate a new paradigm for the next-generation file systems. In it, the traditional view of a file is replaced by the notion of an information unit and the traditional notion of hierarchical filesystems is replaced by an ever-evolving space of dynamically inter-related information units. An Organic File System (OFS) is defined as a system, which develops naturally and which does not conform to artificial rules and predefined, static ways of being viewed by its users. At the core of OFS lie novel abstractions which support a continuously evolving set of information units, users' characterizations of them, and relationships established between them by users accessing them. The abstractional also facilitate the same system and its contents to be viewed differently by different types of users, based on the current information needs. OFS is human-centered, as human input is used to characterize information units and to discover and annotate relationships between units. Given this, at the heart of OFS lie algorithms for content-based search of stored files. We report our R&D efforts so far, including a kernel-level architecture and implementation of the basic features of OFS and relevant performance measures establishing the viability of our approach. We then discuss the large number of challenges that remain and the impact OFS can have in relevant R\&D efforts, highlighting relevant research from other fields, such as Information Retrieval, Social Software, User Interfaces, and Data Management.

CONTEXTO: O Coeficiente de Mortalidade Infantil (CMI), que expressa o risco de um nascido vivo morrer antes de completar um ano de vida, é considerado um dos mais eficientes sensores de desenvolvimento social, econômico e ético, e seu acompanhamento permite inferir sobre a qualidade de vida de uma população. No Rio Grande do Sul, esse coeficiente vem apresentando tendência decrescente, permanecendo abaixo da média nacional. Entretanto, ampliar a compreensão dos determinantes da mortalidade infantil pode contribuir na elaboração de políticas e programas de saúde específicos. São inúmeros os fatores de risco citados na literatura, e a maioria deles é evidenciada em estudos que desconsideram a hierarquia existente nos dados. Porém, crianças que vivem em determinadas regiões podem apresentar características similares, quando comparadas a outras que vivem em regiões diferentes. Assim, as técnicas clássicas de análise, que pressupõem independência entre as observações, podem produzir estimativas viesadas. OBJETIVOS: O objetivo deste estudo foi utilizar os dados de sistemas de informações para analisar a evolução e os determinantes da mortalidade infantil e seus componentes no Rio Grande do Sul, de 1994 a 2004, assim como identificar os fatores associados à mortalidade neonatal, em 2003, considerando características individuais e contextuais. MÉTODO: Para a análise da evolução, foi realizado um estudo ecológico longitudinal, considerando-se medidas repetidas e regressão linear multinível, com microrregiões no nível 2 e tempo no nível 1. Para identificar os determinantes associados ao óbito neonatal, foi utilizada uma coorte retrospectiva que vinculou os nascimentos registrados no período de 01/01/2003 a 03/12/2003 aos óbitos neonatais originados desses nascimentos. Esses fatores foram estimados e comparados por meio da análise dos modelos de regressão logística clássica e multinível. RESULTADOS: Verificou-se que a taxa de mortalidade infantil reduziu de 19,2 para 15,2 por mil nascidos vivos, e as principais causas de óbitos infantis, nos últimos cinco anos, foram as afecções perinatais (54,10%). Aproximadamente 47% da variação nas taxas de mortalidade ocorreu no nível das microrregiões, sendo que 10% de acréscimo na cobertura do Programa Saúde da Família esteve associado à redução de 1‰ na mortalidade infantil, e um acréscimo de 10% na taxa de pobreza esteve associado com uma redução de 2,1‰ nos óbitos infantis. Também, encontrou-se associação positiva com a proporção de baixo peso e a taxa de leitos hospitalares na população e, negativa, com a proporção de partos cesáreos e a taxa de hospitais. As variáveis associadas ao óbito neonatal, no modelo clássico, foram: baixo peso ao nascer, Apgar no 1º e 5º minuto inferiores a 8, presença de anomalia congênita, parto cesáreo, prematuridade e perda fetal anterior. No modelo multinível, essa variável não se manteve significativa, mas a inclusão da variável contextual indicou que 15% da variação da mortalidade neonatal pode ser explicada pela variabilidade nas taxas de pobreza em cada microrregião. CONCLUSÕES: Este estudo evidenciou a predominância dos fatores individuais na mortalidade infantil e neonatal, mas demonstrou que a análise multinível foi capaz de identificar efeitos contextuais, possibilitando ações públicas direcionadas aos grupos vulneráveis. / CONTEXT: The Infant Mortality Coefficient (IMC), that express the risk of a bornalive baby die before completing one year of life, is considered one of the most efficient sensors of social, economic and ethical development, and its following allows to infer on the population life quality. In Rio Grande do Sul this coefficient has presented a decreasing trend, remaining below national average. However, to extend the understanding determinants of infant mortality can contribute in the elaboration of policies and specific health programs. Several risk factors are mentioned in literature, and the majority of them are evidenced in studies that disrespect the existing hierarchy in data. However, children who live in certain regions can present similar characteristics, when compared to others who live in different regions. Thus, classical techniques of analysis that estimate independence between comments, can produce biased estimates. OBJECTIVES: The objective of this study was to use the systems of information data to analyze the evolution and determinants of infant mortality and their components in Rio Grande do Sul from 1994 to 2004, as well as to identify the factors associated to neonatal mortality, in 2003, considering individual and contextual characteristics. METHOD: For the evolution analysis a longitudinal ecologic study was carried out, considering repeated-measures and multilevel linear regression, with microregions in level 2 and time in level 1. To identify the determinants associated to neonatal death, a historic cohort was used to link births recorded from 01/01/2003 to 12/03/2003 with the originated neonatal deaths of these births. These factors were estimated and compared by classic and multilevel logistic regression models. RESULTS: It was verified that the infant mortality rate decreased from 19.2 to 15.2 per thousand live births, and the main causes of infant deaths in the last five years has been perinatal affections (54.10%). Approximately 47% of the variation in mortality rates occurred at a microregion level, being that 10% increase in Family Health Program coverage was associated to the reduction of 1‰ in infant mortality, and an increase of 10% in poverty rate was associated to an increase of 2.1‰ in infant deaths. Also, there was positive association with the proportion of low weight and hospital bed rates in the population and, negative, with the proportion of caesarean sections and hospital rates. Low birthweight, Apgar scores at 1 and at 5 minutes lower 8, presence of congenital abnormality, caesarean section, pre-term birth and previous fetal loss were associated to neonatal deaths in the classical model. In the multilevel model, previous fetal loss did not remain significant, but the inclusion of contextual variable indicated that 15% of neonatal mortality variation can be explained by the variability in rates of poverty in each microregion. CONCLUSIONS: This study evidenced the predominance of individual factors in infant and neonatal mortality, but it demonstrated that the multilevel analysis was capable of identifying contextual effects, making directed actions to the susceptible groups possible.

CONTEXTO: O Coeficiente de Mortalidade Infantil (CMI), que expressa o risco de um nascido vivo morrer antes de completar um ano de vida, é considerado um dos mais eficientes sensores de desenvolvimento social, econômico e ético, e seu acompanhamento permite inferir sobre a qualidade de vida de uma população. No Rio Grande do Sul, esse coeficiente vem apresentando tendência decrescente, permanecendo abaixo da média nacional. Entretanto, ampliar a compreensão dos determinantes da mortalidade infantil pode contribuir na elaboração de políticas e programas de saúde específicos. São inúmeros os fatores de risco citados na literatura, e a maioria deles é evidenciada em estudos que desconsideram a hierarquia existente nos dados. Porém, crianças que vivem em determinadas regiões podem apresentar características similares, quando comparadas a outras que vivem em regiões diferentes. Assim, as técnicas clássicas de análise, que pressupõem independência entre as observações, podem produzir estimativas viesadas. OBJETIVOS: O objetivo deste estudo foi utilizar os dados de sistemas de informações para analisar a evolução e os determinantes da mortalidade infantil e seus componentes no Rio Grande do Sul, de 1994 a 2004, assim como identificar os fatores associados à mortalidade neonatal, em 2003, considerando características individuais e contextuais. MÉTODO: Para a análise da evolução, foi realizado um estudo ecológico longitudinal, considerando-se medidas repetidas e regressão linear multinível, com microrregiões no nível 2 e tempo no nível 1. Para identificar os determinantes associados ao óbito neonatal, foi utilizada uma coorte retrospectiva que vinculou os nascimentos registrados no período de 01/01/2003 a 03/12/2003 aos óbitos neonatais originados desses nascimentos. Esses fatores foram estimados e comparados por meio da análise dos modelos de regressão logística clássica e multinível. RESULTADOS: Verificou-se que a taxa de mortalidade infantil reduziu de 19,2 para 15,2 por mil nascidos vivos, e as principais causas de óbitos infantis, nos últimos cinco anos, foram as afecções perinatais (54,10%). Aproximadamente 47% da variação nas taxas de mortalidade ocorreu no nível das microrregiões, sendo que 10% de acréscimo na cobertura do Programa Saúde da Família esteve associado à redução de 1‰ na mortalidade infantil, e um acréscimo de 10% na taxa de pobreza esteve associado com uma redução de 2,1‰ nos óbitos infantis. Também, encontrou-se associação positiva com a proporção de baixo peso e a taxa de leitos hospitalares na população e, negativa, com a proporção de partos cesáreos e a taxa de hospitais. As variáveis associadas ao óbito neonatal, no modelo clássico, foram: baixo peso ao nascer, Apgar no 1º e 5º minuto inferiores a 8, presença de anomalia congênita, parto cesáreo, prematuridade e perda fetal anterior. No modelo multinível, essa variável não se manteve significativa, mas a inclusão da variável contextual indicou que 15% da variação da mortalidade neonatal pode ser explicada pela variabilidade nas taxas de pobreza em cada microrregião. CONCLUSÕES: Este estudo evidenciou a predominância dos fatores individuais na mortalidade infantil e neonatal, mas demonstrou que a análise multinível foi capaz de identificar efeitos contextuais, possibilitando ações públicas direcionadas aos grupos vulneráveis. / CONTEXT: The Infant Mortality Coefficient (IMC), that express the risk of a bornalive baby die before completing one year of life, is considered one of the most efficient sensors of social, economic and ethical development, and its following allows to infer on the population life quality. In Rio Grande do Sul this coefficient has presented a decreasing trend, remaining below national average. However, to extend the understanding determinants of infant mortality can contribute in the elaboration of policies and specific health programs. Several risk factors are mentioned in literature, and the majority of them are evidenced in studies that disrespect the existing hierarchy in data. However, children who live in certain regions can present similar characteristics, when compared to others who live in different regions. Thus, classical techniques of analysis that estimate independence between comments, can produce biased estimates. OBJECTIVES: The objective of this study was to use the systems of information data to analyze the evolution and determinants of infant mortality and their components in Rio Grande do Sul from 1994 to 2004, as well as to identify the factors associated to neonatal mortality, in 2003, considering individual and contextual characteristics. METHOD: For the evolution analysis a longitudinal ecologic study was carried out, considering repeated-measures and multilevel linear regression, with microregions in level 2 and time in level 1. To identify the determinants associated to neonatal death, a historic cohort was used to link births recorded from 01/01/2003 to 12/03/2003 with the originated neonatal deaths of these births. These factors were estimated and compared by classic and multilevel logistic regression models. RESULTS: It was verified that the infant mortality rate decreased from 19.2 to 15.2 per thousand live births, and the main causes of infant deaths in the last five years has been perinatal affections (54.10%). Approximately 47% of the variation in mortality rates occurred at a microregion level, being that 10% increase in Family Health Program coverage was associated to the reduction of 1‰ in infant mortality, and an increase of 10% in poverty rate was associated to an increase of 2.1‰ in infant deaths. Also, there was positive association with the proportion of low weight and hospital bed rates in the population and, negative, with the proportion of caesarean sections and hospital rates. Low birthweight, Apgar scores at 1 and at 5 minutes lower 8, presence of congenital abnormality, caesarean section, pre-term birth and previous fetal loss were associated to neonatal deaths in the classical model. In the multilevel model, previous fetal loss did not remain significant, but the inclusion of contextual variable indicated that 15% of neonatal mortality variation can be explained by the variability in rates of poverty in each microregion. CONCLUSIONS: This study evidenced the predominance of individual factors in infant and neonatal mortality, but it demonstrated that the multilevel analysis was capable of identifying contextual effects, making directed actions to the susceptible groups possible.