Tesis por compendio / [ES] El cáncer es la segunda causa de muerte en el mundo y se caracteriza principalmente por la proliferación descontrolada de las células que forman el tumor. Aunque el desarrollo de un tumor es posible debido a ciertos procesos comunes desencadenados por la desregulación del equilibrio existente entre los componentes moleculares de una célula y sus elementos de control, existe una gran heterogeneidad en los mecanismos a través de los cuales ocurre dicha desregulación. Gracias al desarrollo de nuevas tecnologías de secuenciación ha sido posible observar como esta heterogeneidad no solo se observa entre los distintos tipos de tumores sino entre las propias células de un mismo tumor. La caracterización de la heterogeneidad tumoral ha tenido un gran impacto en la comprensión de la enfermedad y el desarrollo de nuevas terapias dirigidas. Por este motivo, con el fin de mejorar la caracterización de alteraciones en los distintos mecanismos regulatorios, en esta tesis se han desarrollado dos metodologías con gran potencial para su aplicación en la medicina personalizada y que permiten estudiar la heterogeneidad inter e intratumoral de los estados de activación de elementos reguladores. En primer lugar, se desarrolló una metodología que permite determinar en una muestra el estado de activación de los factores de transcripción (FTs) a partir de la expresión de los genes a los que regula. Se aplicó la metodología para realizar un análisis sistemático de varios cánceres (conocido como estudios pan-cáncer) en el que se caracterizó por primera vez el escenario regulatorio de 52 FTs en 11 tipos de cáncer distintos. Además, al poder obtener valores de activación individuales para cada muestra, fue posible observar correlaciones entre la activación de algunos FTs con la supervivencia, sugiriendo así su uso como marcadores pronósticos. En segundo lugar, se desarrolló otra metodología en la que se emplea un modelo mecanístico para determinar el estado de activación de alrededor de 1000 circuitos de señalización a partir de datos de experimentos transcriptómicos de células únicas (scRNAseq). El uso de este modelo mecanístico en datos de scRNAseq de 4 pacientes de glioblastoma, además de mostrar la heterogeneidad intratumoral presente en las muestras, ha permitido realizar intervenciones in silico para simular el efecto de distintas drogas sobre las células. De esta manera, ha sido posible describir posibles mecanismos mediante los cuales un grupo de células pueden evitar el efecto de una terapia dirigida. Las metodologías desarrolladas en esta tesis, así como los resultados obtenidos tras su aplicación supone una valiosa fuente de información para el desarrollo de marcadores de diagnóstico, pronóstico y respuesta que ayuden a entender mejor los distintos niveles de heterogeneidad presentes en cáncer, y así, poder aumentar la eficacia de las terapias dirigidas. / [CA] El càncer és la segona causa de mort al món i es caracteritza principalment per la proliferació descontrolada de les cèl·lules que formen el tumor. Encara que el desenvolupament d'un tumor és possible a causa de certs processos comuns desencadenats per la desregulació de l'equilibri existent entre els components moleculars d'una cèl·lula i els seus elements de control, hi ha una gran heterogeneïtat en els mecanismes a través dels quals s'aconseguix aquesta desregulació. Gràcies a el desenvolupament de noves tecnologies de seqüenciació ha sigut possible observar com aquesta heterogeneïtat no només s'observa entre els diferents tipus de tumors sinó entre les pròpies cèl·lules d'un mateix tumor. La caracterització de l'heterogeneïtat tumoral ha tingut un gran impacte en la comprensió de la malaltia i el desenvolupament de noves teràpies dirigides. Per aquest motiu, per tal de millorar la caracterització d'alteracions en els diferents mecanismes reguladors, en aquesta tesi s'han desenvolupat dues metodologies amb gran potencial per a la seua aplicació en la medicina personalitzada i que permeten estudiar l'heterogeneïtat inter i intratumoral dels estats de activació d'elements reguladors. En primer lloc es va desenvolupar una metodologia que permet determinar en una mostra l'estat d'activació dels factors de transcripció (FTs) a partir de l'expressió dels gens als que regula. Es va aplicar la metodologia per a realitzar una anàlisi de pan-cancer en el qual es va caracteritzar per primera vegada l'escenari regulatori de 52 FTs a 11 tipus de càncer diferents. A més, al poder obtenir valors d'activació individuals per a cada mostra, va ser possible observar correlacions entre l'activació d'alguns FTs amb la supervivència, suggerint així el seu ús com a marcadors pronòstics. En segon lloc, es va desenvolupar una altra metodologia en la qual s'empra un model mecanístic per determinar l'estat d'activació d'al voltant de 1000 circuits de senyalització a partir d'experiments transcriptòmics de cèl·lules úniques (scRNAseq). L'ús d'aquest model mecanístic en dades de scRNAseq de 4 pacients de glioblastoma, a més de mostrar l'heterogeneïtat intratumoral present en les mostres, ha permès realitzar intervencions in silico per simular l'efecte de diferents drogues sobre les cèl·lules. D'aquesta manera, ha estat possible descriure possibles mecanismes mitjançant els quals un grup de cèl·lules poden evitar l'efecte d'una teràpia dirigida. Les metodologies desenvolupades en aquesta tesi, així com els resultats obtinguts després de la seva aplicació suposa una valuosa font d'informació per al desenvolupament de marcadors de diagnòstic, pronòstic i resposta que ajudin a entendre millor els diferents nivells d'heterogeneïtat presents en càncer, i així, poder augmentar l'eficàcia de les teràpies dirigides. / [EN] Cancer is the second leading cause of death in the world and is characterized mainly by the uncontrolled proliferation of the cells that make up the tumor. Although the development of a tumor is possible due to certain common processes triggered by the dysregulation of the existing balance between the molecular components of a cell and its control elements, there is great heterogeneity in the mechanisms through which this dysregulation is achieved. Thanks to the development of new sequencing technologies, it has been possible to observe how this heterogeneity is not only observed between the different types of tumors but also between the cells of the same tumor. The characterization of tumor heterogeneity has had a great impact on the understanding of the disease and the development of new targeted therapies. For this reason, in order to improve the characterization of alterations in the different regulatory mechanisms, in this thesis two methodologies have been developed that allow studying the inter- and intratumoral heterogeneity of the activation states of regulatory elements and with great potential for their application in personalized medicine. In the first place, a methodology that allows determining in a sample the activation state of the transcription factors (FTs) from the expression of the genes that it regulates was developed. The methodology was applied to perform a pan-cancer analysis in which the regulatory scenario of 52 FTs was characterized for the first time in 11 different types of cancer. Furthermore, by being able to obtain individual activation values for each sample, it was possible to observe correlations between the activation of some FTs with survival, thus suggesting their use as prognostic markers. Second, another methodology was developed using a mechanistic model to determine the activation state of around 1000 signaling circuits in single cell transcriptomic experiments (scRNAseq). The use of this mechanistic model in scRNAseq data from 4 glioblastoma patients, in addition to showing the intratumoral heterogeneity present in the samples, has allowed in silico interventions to simulate the effect of different drugs on cells. In this way, it has been possible to describe possible mechanisms by which a group of cells can avoid the effect of a targeted therapy. The methodologies developed in this thesis, as well as the results obtained after its application, is a valuable source of information for the development of diagnostic, prognostic and response markers that help to better understand the different levels of heterogeneity present in cancer, and thus, be able increase the effectiveness of targeted therapies. / Marín Falco, M. (2021). Estudio de la heterogeneidad regulatoria en cáncer y sus implicaciones en la medicina personalizada [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/165413 / Compendio

Yes / Fused deposition modelling (FDM) 3D printing has shown the most immediate potential for on-demand dose personalisation to suit particular patient's needs. However, FDM 3D printing often involves employing a relatively large molecular weight thermoplastic polymer and results in extended release pattern. It is therefore essential to fast-track drug release from the 3D printed objects. This work employed an innovative design approach of tablets with unique built-in gaps (Gaplets) with the aim of accelerating drug release. The novel tablet design is composed of 9 repeating units (blocks) connected with 3 bridges to allow the generation of 8 gaps. The impact of size of the block, the number of bridges and the spacing between different blocks was investigated. Increasing the inter-block space reduced mechanical resistance of the unit, however, tablets continued to meet pharmacopeial standards for friability. Upon introduction into gastric medium, the 1 mm spaces gaplet broke into mini-structures within 4 min and met the USP criteria of immediate release products (86.7% drug release at 30 min). Real-time ultraviolet (UV) imaging indicated that the cellulosic matrix expanded due to swelling of hydroxypropyl cellulose (HPC) upon introduction to the dissolution medium. This was followed by a steady erosion of the polymeric matrix at a rate of 8 μm/min. The design approach was more efficient than a comparison conventional formulation approach of adding disintegrants to accelerate tablet disintegration and drug release. This work provides a novel example where computer-aided design was instrumental at modifying the performance of solid dosage forms. Such an example may serve as the foundation for a new generation of dosage forms with complicated geometric structures to achieve functionality that is usually achieved by a sophisticated formulation approach.

Cette recherche définit la personnalisation comme une approche centrée sur l'apprenant où la personne est prise dans son ensemble (émotions, cognition et socialisation). L’apprenant est actif, prend des initiatives et collabore avec les autres. Elle détaille les attributs de l'apprentissage personnalisé, ses perspectives et ses théories et propose un modèle personnalisé testé via le développement d’une application mobile pour les cours de base en informatique pour les étudiants en première année universitaire. Les résultats portent principalement sur l'impact de la personnalisation sur la performance, l'engagement et l'autonomie des apprenants. Les analyses de données montrent que la performance du groupe personnalisé dans les tests théoriques est significativement meilleure que les groupes suivant un enseignement traditionnel, mais les résultats ne sont pas significatifs pour les tests pratiques. La performance des apprenants dans le groupe personnalisé est indépendante du type d'évaluation choisi (tests théoriques et pratiques, projets individuels ou collaboratifs). Les apprenants ont manifesté une certaine responsabilité et un comportement autonome et ils ont été capables, avec une grille,d’évaluer d’une manière correcte et objective le travail de leurs pairs. L'engagement émotionnel est le type d'engagement le plus élevé et il apparaît indépendant de l’engagement cognitif, comportemental et agentique qui le suivent dans l’ordre respectif. / This research adopts a definition of personalization as a learner centered approach that takes into consideration the person as a whole (emotions, cognition and socialization) and where the learner is active, takes initiatives and collaborates with others. It details attributes of personalized learning, its perspectives and theories behind it. It proposes a personalized model that is tested through designing a mobile app for computer basics courses for first year university students. Results focus mainly on the impact of personalization on learners’ performance, engagement and autonomy. Data analyses showed that the personalized group did significantly better in theoretical tests than groups taught in a traditional way, but results were not significant for practical tests. The performance of learners in the personalized group does not depend on the type of evaluation used (theoretical and practical tests, individual or collaborative projects). Learners manifested autonomous behavior and responsibility and were able with a rubric in hand to conduct a fair and objective judgement of peers’ work. The emotional engagement was the highest type of engagement and was independent of the cognitive, behavioral and agentic one that come next respectively.

Purpose: The aim of this paper is to explain and to create an understanding if personalized advertising online creates value for customers.  Design/methodology/approach: A qualitative study through 14 semi-structured interviews.  Findings: The study found personalized advertising to be seen as value co-creation in some cases, but because privacy concerns exist and affect the perception of advertising, it can in many cases lead to value co-destruction instead. It is thus a consideration between privacy concerns and the perceived value of the personalized advertising that decides if the offering will co-create or co-destroy value.  Research limitations/implications for future research: Our study did not involve respondents’ younger than 21 years old, which could have affected the result as this is a generation seen as technology savvy. Through a quantitative study, future research could try to find extremes in personalities by conducting a survey with a large sample of people in different ages, nationalities, gender, active online, etc. in order to see if there are correlations between for example age and privacy concerns.  Practical implications: One purpose of the study is to provide companies with insights of how different customers perceive personalized advertising online in terms of customer value in order for companies to know how to think when targeting their customers.  Keywords: online advertising, personalized advertising, personalized-privacy paradox, privacy concerns, value creation, value co-creation, value co-destruction.

A Gamificação é uma técnica que a utiliza elementos de design de jogos em ambientes que não são jogos, visando aumentar a motivação e engajamento dos usuários e que vem ganhando espaço em diversos áreas como saúde, marketing e também na educação. Porém, ainda que o interesse pela técnica venha crescendo, os meios para sua aplicação nesses ainda não estão bem definidos e os resultados obtidos têm-se mostrado dependentes do contexto e da população alvo. Diversos autores atribuem essa inconstância nos resultados a problemas no design da gamificação, uma vez que a maior parte dos projetos tem utilizado abordagens one-size-fits-all, no qual todos os usuários utilizam o mesmo ambiente independente de suas preferências individuais. Diante desse cenário, tem-se proposto que a gamificação personalizada pode atender uma maior parcela dos usuários, adequando os ambientes gamificados ao perfil dos usuários. Uma das abordagens para a personalização da gamificação consiste no uso de tipologias de jogadores para determinar os elementos mais interessante para o usuário. No entanto, as tipologias utilizam estereótipos, criando constructos que ainda restringem as informações consideradas durante a personalização. Dessa forma, neste trabalho buscou-se investigar a personalização com base na teoria de motivações para se engajarem em jogos, que trata o perfil do usuário como um conjunto de diferentes subcomponentes motivacionais correlacionados, que se agrupam em macro-componentes. Para isso, adaptou-se a teoria para o contexto da gamificação e elaborou-se dois modelos o de Macro-Gamificação, o qual relaciona-se com a teoria de Autodeterminação e às necessidades de Competência, Relacionamento e Autonomia do usuário, e o de Micro-Gamificação, que relaciona os elementos de jogos a um determinado subcomponente motivacional e disponibilizá-lo mediante o interesse do usuário no subcomponente. Para avaliar então se a gamificação personalizada influencia no engajamento dos usuários quando comparada a gamificação não personalizada, os modelos foram implementados em um ambiente virtual de aprendizagem, preparado para criar os perfis de gamificação dos usuários dinamicamente e adaptar interface do em tempo real. Realizou-se então um estudo de caso com N=36, utilizando como domínio o estudo dos silabários do idioma japonês. Ao final do estudo foram identificados dois padrões de atuação no sistema com uma diferença de 65% de participação e que foi utilizado para segmentar os participantes. No segmento menos engajado, os participantes do grupo não personalizado apresentaram um engajamento aos grupos personalizados. Já no segmento dos usuários mais ativos o grupo utilizando o modelo Micro-Gamificado, apresentou-se mais engajado. Desse modo, não é possível afirmar que a gamificação personalizada proporcione um maior engajamento do que a gamificação sem personalização, embora os resultados sugiram que usuários que permanecem utilizando o sistema por mais tempo tem um maior engajamento no ambiente personalizado. Por fim, é possível afirmar que o desenvolvimento de sistemas gamificados com personalização ainda está em sua infância e por isso nesta pesquisa além de buscar evidencias sobre o impacto da gamificação personalizada no engajamento dos usuários, buscou-se também desenvolver ferramental para facilitar o processo para os membros da comunidade em ordem de impulsionar os avanços dessa área de pesquisa. / Gamification is a technique that uses game design elements in non-game context, to increase users motivation and engagement and that has been gaining space in several areas such as health, marketing and also in education. However, although the interest in the technique is growing, the means for its application are still not well defined and the results obtained have been shown to be dependent on the context and the population. Several authors attribute this resultsin the results to problems in gamification design, since most projects have been using an one-size-fitsall approach, in which all users uses the same environment independent of their preferences. Given this scenario, it has been proposed that the personalized gamification can adress a larger portion of users, adapting the gamified environments to users profiles One of the approaches to personalize the gamification is to use player typologies to determine which elements are most interesting to the user. However, typologies uses stereotypes, creating constructs that still restrict the information considered during customization. Thus, in this work, we sought to investigate personalization based on the theory of motivations to engage in games, which treats the user profile as a set of different correlated motivational subcomponents, which are grouped into macrocomponents. For this, the theory was adapted to the context of the gamification and two models were elaborated the Macro-Gamification, which is related to the theory of Self-determination and to the needs of Competence, Relationship and Autonomy of the user, and the Micro-Gamification, which relates the game elements to a particular motivational subcomponent and make it available through the users interest in the subcomponent. In order to evaluate whether personalized gamification influences user engagement when compared to non-personalized gamification, the models were implemented in a virtual learning environment, prepared to dynamically create users gamification profiles and adapt the interface in real time. A case study was then carried out with N = 36, using as a domain the study of syllabaries of the Japanese language. At the end of the study, two patterns of performance in the system with a difference of 65 % participation were identified and used to segment the participants. In the less engaged segment, the non-personalized group participants showed a higer engagement than the personalized groups. However, in the segment of the most active users the group using the Micro-Gamified model, presented itself more engaged. Thus, it can not be argued that personalized gamification provides greater engagement than non-personalized gamification, although the results suggest that users who remain using the system longer have a greater engagement in the personalized approach. Finally, it is possible to affirm that the development of personalized gamified systems is still in its infancy and for this reason, in this research, besides searching for evidence on the impact of personalized gamification on user engagement, we also sought to develop tooling to facilitate the process for the members of the community in order to boost the advances of this area of research.

A Gamificação é uma técnica que a utiliza elementos de design de jogos em ambientes que não são jogos, visando aumentar a motivação e engajamento dos usuários e que vem ganhando espaço em diversos áreas como saúde, marketing e também na educação. Porém, ainda que o interesse pela técnica venha crescendo, os meios para sua aplicação nesses ainda não estão bem definidos e os resultados obtidos têm-se mostrado dependentes do contexto e da população alvo. Diversos autores atribuem essa inconstância nos resultados a problemas no design da gamificação, uma vez que a maior parte dos projetos tem utilizado abordagens one-size-fits-all, no qual todos os usuários utilizam o mesmo ambiente independente de suas preferências individuais. Diante desse cenário, tem-se proposto que a gamificação personalizada pode atender uma maior parcela dos usuários, adequando os ambientes gamificados ao perfil dos usuários. Uma das abordagens para a personalização da gamificação consiste no uso de tipologias de jogadores para determinar os elementos mais interessante para o usuário. No entanto, as tipologias utilizam estereótipos, criando constructos que ainda restringem as informações consideradas durante a personalização. Dessa forma, neste trabalho buscou-se investigar a personalização com base na teoria de motivações para se engajarem em jogos, que trata o perfil do usuário como um conjunto de diferentes subcomponentes motivacionais correlacionados, que se agrupam em macro-componentes. Para isso, adaptou-se a teoria para o contexto da gamificação e elaborou-se dois modelos o de Macro-Gamificação, o qual relaciona-se com a teoria de Autodeterminação e às necessidades de Competência, Relacionamento e Autonomia do usuário, e o de Micro-Gamificação, que relaciona os elementos de jogos a um determinado subcomponente motivacional e disponibilizá-lo mediante o interesse do usuário no subcomponente. Para avaliar então se a gamificação personalizada influencia no engajamento dos usuários quando comparada a gamificação não personalizada, os modelos foram implementados em um ambiente virtual de aprendizagem, preparado para criar os perfis de gamificação dos usuários dinamicamente e adaptar interface do em tempo real. Realizou-se então um estudo de caso com N=36, utilizando como domínio o estudo dos silabários do idioma japonês. Ao final do estudo foram identificados dois padrões de atuação no sistema com uma diferença de 65% de participação e que foi utilizado para segmentar os participantes. No segmento menos engajado, os participantes do grupo não personalizado apresentaram um engajamento aos grupos personalizados. Já no segmento dos usuários mais ativos o grupo utilizando o modelo Micro-Gamificado, apresentou-se mais engajado. Desse modo, não é possível afirmar que a gamificação personalizada proporcione um maior engajamento do que a gamificação sem personalização, embora os resultados sugiram que usuários que permanecem utilizando o sistema por mais tempo tem um maior engajamento no ambiente personalizado. Por fim, é possível afirmar que o desenvolvimento de sistemas gamificados com personalização ainda está em sua infância e por isso nesta pesquisa além de buscar evidencias sobre o impacto da gamificação personalizada no engajamento dos usuários, buscou-se também desenvolver ferramental para facilitar o processo para os membros da comunidade em ordem de impulsionar os avanços dessa área de pesquisa. / Gamification is a technique that uses game design elements in non-game context, to increase users motivation and engagement and that has been gaining space in several areas such as health, marketing and also in education. However, although the interest in the technique is growing, the means for its application are still not well defined and the results obtained have been shown to be dependent on the context and the population. Several authors attribute this resultsin the results to problems in gamification design, since most projects have been using an one-size-fitsall approach, in which all users uses the same environment independent of their preferences. Given this scenario, it has been proposed that the personalized gamification can adress a larger portion of users, adapting the gamified environments to users profiles One of the approaches to personalize the gamification is to use player typologies to determine which elements are most interesting to the user. However, typologies uses stereotypes, creating constructs that still restrict the information considered during customization. Thus, in this work, we sought to investigate personalization based on the theory of motivations to engage in games, which treats the user profile as a set of different correlated motivational subcomponents, which are grouped into macrocomponents. For this, the theory was adapted to the context of the gamification and two models were elaborated the Macro-Gamification, which is related to the theory of Self-determination and to the needs of Competence, Relationship and Autonomy of the user, and the Micro-Gamification, which relates the game elements to a particular motivational subcomponent and make it available through the users interest in the subcomponent. In order to evaluate whether personalized gamification influences user engagement when compared to non-personalized gamification, the models were implemented in a virtual learning environment, prepared to dynamically create users gamification profiles and adapt the interface in real time. A case study was then carried out with N = 36, using as a domain the study of syllabaries of the Japanese language. At the end of the study, two patterns of performance in the system with a difference of 65 % participation were identified and used to segment the participants. In the less engaged segment, the non-personalized group participants showed a higer engagement than the personalized groups. However, in the segment of the most active users the group using the Micro-Gamified model, presented itself more engaged. Thus, it can not be argued that personalized gamification provides greater engagement than non-personalized gamification, although the results suggest that users who remain using the system longer have a greater engagement in the personalized approach. Finally, it is possible to affirm that the development of personalized gamified systems is still in its infancy and for this reason, in this research, besides searching for evidence on the impact of personalized gamification on user engagement, we also sought to develop tooling to facilitate the process for the members of the community in order to boost the advances of this area of research.

This study examines how smartphones, laptops and tablets are used to create a sense of security, place, and time in the everyday lives of their consumers. Mark Deuze (2012) and his take on the modern society as a zombie apocalypse has been an inspiration for my work, there of the title. ”We’re all fucking zombies” is a metaphor for how highly-connected people of today are by living in and through the media, instead of living with it. To implement this I’ve chosen a phenomenological perspective, as it is the media-users’ subjective experiences of their everyday lives that I’m mainly interested in studying. I decided to focus on the mobile use of media since most of the research in this area focus on domestic media use. The theoretical framework that has set the foundation for the study is a combination of the time- space-dimensions of mobility, media as practice, symbolic interactionism, relational artifacts and phenomenological sociology. The purpose of using these theories is to get insight on how the media creates new opportunities for our social life, and to get an overview of how the new technological media leads to entirely new types of practices. The empirical data has been collected through a qualitative focus group interview with four respondents. They were selected to participate as they perceive themselves to be above average in comparison with the statistics of Mediebarometern (2014). The results showed that being connected through mobile media is considered to be of high importance. Although the ”connectedness” is only vital when being present in the locations directly related to the everyday life. Based on Silverstones (1994) explanation of phenomenology, and his studies of how television contributes to the ontological security, I conclude that the personalized mobile media has the same effect on it’s users.

With new treatments and novel technology available, personalized medicine has become a key topic in the new era of healthcare. Traditional statistical methods for personalized medicine and subgroup identification primarily focus on single treatment or two arm randomized control trials (RCTs). With restricted inclusion and exclusion criteria, data from RCTs may not reflect real world treatment effectiveness. However, electronic medical records (EMR) offers an alternative venue. In this paper, we propose a general framework to identify individualized treatment rule (ITR), which connects the subgroup identification methods and ITR. It is applicable to both RCT and EMR data. Given the large scale of EMR datasets, we develop a recursive partitioning algorithm to solve the problem (ITR-Tree). A variable importance measure is also developed for personalized medicine using random forest. We demonstrate our method through simulations, and apply ITR-Tree to datasets from diabetes studies using both RCT and EMR data. Software package is available at https://github.com/jinjinzhou/ITR.Tree.

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.

鑑於國人網路依賴度越來越高，每人每日透過網路獲取豐富的資訊及資料，然而由於資訊過多也造成使用者不知道自己真正想要的資訊，因此找尋出個人化資訊勢必成為未來研究的方向。一般網路行銷廣告都是在分析使用者在網站上的行為，然而鮮少去分析到使用者端環境及儲存的資料。相較於使用者在網站上的行為，使用者端所留下的行為資訊更能夠反應出使用者的真正興趣，因此本研究主要探討在個人使用環境(如：PC、NB)中，透過使用者最常接觸的三種途徑(包括：上網瀏覽資訊、信件資訊及常用檔案內容資訊)來獲取使用者興趣的資訊，並且建構出一套個人化廣告推薦系統常駐於使用者端即時(real-time)記錄使用者的行為資訊。本系統運用推薦技術(Recommendation Technique)搭配關聯式法則(Association Rule)來將這些資訊有效的過濾並關聯出使用者的喜好及興趣，同時利用這些資訊上網找尋合適的廣告資料，用以建構出個人化廣告推薦模式。 / The rapid growth of Internet has changed the patterns of our life. Everyone can gain rich information on internet, but plenty of information will confuse user's ability to determine whether these information are useful. Therefore, the further trend is to discover personalized information. Many researches about internet marketing advertisement are mining user's behavior in website server, but scarce researches focus on client. Compared to user's behavior in website, the behavior information which stays in local environment (ex. PC, NB) can reflect user's profile more. Thus, this research mainly discuss how to record user's behavior information containing Web Page Title, E-mail Subject and Document Content in local environment and how to construct a personalized advertisement recommendation system resident in memory of local environment for timely (On-line) collecting user's behavior information to create “user profile” using recommendation technique and association rule. This system will utilize “user profile” to provide appropriate personalized advertisement for user. Finally, we apply several experiments to verify the feasibility of our system.