Ontologies are a key factor of Information management as they provide a common representation to any domain. Historically, finance domain has suffered from a lack of efficiency in managing vast amounts of financial data, a lack of communication and knowledge sharing between analysts. Particularly, with the growth of fraud in financial markets, cases are challenging, complex, and involve a huge volume of information. Gathering facts and evidence is often complex. Thus, the impetus for building a financial fraud ontology arises from the continuous improvement and development of financial market surveillance systems with high analytical capabilities to capture frauds which is essential to guarantee and preserve an efficient market.This thesis proposes an ontology-based approach for financial market surveillance systems. The proposed ontology acts as a semantic representation of mining concepts from unstructured resources and other internet sources (corpus). The ontology contains a comprehensive concept system that can act as a semantically rich knowledge base for a market monitoring system. This could help fraud analysts to understand financial fraud practices, assist open investigation by managing relevant facts gathered for case investigations, providing early detection techniques of fraudulent activities, developing prevention practices, and sharing manipulation patterns from prosecuted cases with investigators and relevant users. The usefulness of the ontology will be evaluated through three case studies, which not only help to explain how manipulation in markets works, but will also demonstrate how the ontology can be used as a framework for the extraction process and capturing information related to financial fraud, to improve the performance of surveillance systems in fraud monitoring. Given that most manipulation cases occur in the unregulated markets, this thesis uses a sample of fraud cases from the unregulated markets. On the empirical side, the thesis presents examples of novel applications of text-mining tools and data-processing components, developing off-line surveillance systems that are fully working prototypes which could train the ontology in the most recent manipulation techniques.

Diploma thesis deals with the ADS-B system and its possible use at the Airport Václav Havel in Prague. In the beginning of this thesis there is described the history leading to the introduction of this system and then the ADS-B is explained. In this thesis is also included a schedule for the introduction of the system. Them the application is presented to the airport with integrity verification, functionality, reliability and safety.

The Internet of Things is a constantly developing field. With advancements of algorithms for object detection and classification for images and videos, the possibilities of what can be made with small and cost efficient edge-devices are increasing. This work presents how camera traps and deep learning can be utilized for surveillance in remote environments, such as animal sanctuaries in the African Savannah. The camera traps connect to a smart surveillance network where images and sensor-data are analysed. The analysis can then be used to produce valuable information, such as the location of endangered animals or unauthorized humans, to park rangers working to protect the wildlife in these animal sanctuaries. Different motion detection algorithms are tested and evaluated based on related research within the subject. The work made in this thesis builds upon two previous theses made within Project Ngulia. The implemented surveillance system in this project consists of camera sensors, a database, a REST API, a classification service, a FTP-server and a web-dashboard for displaying sensor data and resulting images. A contribution of this work is an end-to-end smart surveillance system that can use different camera sources to produce valuable information to stakeholders. The camera software developed in this work is targeting the ESP32 based M5Stack Timer Camera and runs a motion detection algorithm based on Self-Organizing Maps. This improves the selection of data that is fed to the image classifier on the server. This thesis also contributes with an algorithm for doing iterative image classifications that handles the issues of objects taking up small parts of an image, making them harder to classify correctly.

Despite technological advances in public health informatics, the evaluation of infectious disease surveillance systems data remains incomplete. In this study, a thorough evaluation was performed of the West Virginia Electronic Disease Surveillance System (WVEDSS, 2007-2010) and the West Virginia Electronic Disease Surveillance System NationalEDSS -Based System (WVEDSS-NBS; March 2012 - March 2014) for Category II infectious diseases in West Virginia. The purpose was to identify key areas in the surveillance system process from disease diagnosis to disease prevention that need improvement. Grounded in the diffusion of innovation theory, a quasi-experimental, interrupted, time-series design was used to evaluate the 2 data sets. Research questions examined differences in mean reporting time, the 24-hour standard, and comparison of complete fields (DOB, gender etc.) of the data sets using independent samples t tests. The study found (a) that the mean reporting times were shorter for WVEDSS compared to WVEDSS-NBS (p < .05) for all vaccine-preventable infectious diseases (VPID) in Category II except for mumps; (b) that the 24-hour standard was not met for WVEDSS compared to WVEDSS-NBS (p < .05) for all VPID in Category II except for mumps, and (c) that most fields were complete for WVEDSS compared to WVEDSS-NBS (p < .05) for all VPID in Category II except for meningococcal disease. Healthcare professionals in the state can use the results of this research to improve the system attributes of timeliness and completeness. Implications for positive social change included improved access to public health data to better understand health disparities, which, in turn could reduce morbidity and mortality within the population.

La France connaît un vieillissement de sa population sans précédent. La part des séniors s’accroît et notre société se doit de repenser son organisation pour tenir compte de ce changement et mieux connaître cette population.De nombreuses cohortes de personnes âgées existent déjà à travers le monde dont quatre en France et, bien que la part de cette population vivant dans des structures d’hébergement collectif (EHPAD, cliniques de soins de suite) augmente, la connaissance de ces seniors reste lacunaire.Aujourd’hui les groupes privés de maisons de retraite et d’établissements sanitaires comme Korian ou Orpéa s’équipent de grandes bases de données relationnelles permettant d’avoir de l’information en temps réel sur leurs patients/résidents. Depuis 2010 les dossiers de tous les résidents Korian sont dématérialisés et accessibles par requêtes. Ils comprennent à la fois des données médico-sociales structurées décrivant les résidents et leurs traitements et pathologies, mais aussi des données textuelles explicitant leur prise en charge au quotidien et saisies par le personnel soignant.Au fil du temps et alors que le dossier résident informatisé (DRI) avait surtout été conçu comme une application de gestion de base de données, il est apparu comme une nécessité d’exploiter cette mine d’informations et de construire un outil d’aide à la décision destiné à améliorer l’efficacité des soins. L’Institut du Bien Vieillir IBV devenu entretemps la Fondation Korian pour le Bien Vieillir a alors choisi, dans le cadre d’un partenariat Public/Privé de financer un travail de recherche destiné à mieux comprendre le potentiel informatif de ces données, d’évaluer leur fiabilité et leur capacité à apporter des réponses en santé publique. Ce travail de recherche et plus particulièrement cette thèse a alors été pensée en plusieurs étapes.- D’abord l’analyse de contenu du data warehouse DRI, l’objectif étant de construire une base de données recherche, avec un versant social et un autre de santé. Ce fut le sujet du premier article.- Ensuite, par extraction directe des informations socio-démographiques des résidents dès leur entrée, de leurs hospitalisations et décès puis, par un processus itératif d’extractions d’informations textuelles de la table des transmissions et l’utilisation de la méthode Delphi, nous avons généré vingt-quatre syndromes, ajouté les hospitalisations et les décès et construit une base de données syndromique, la Base du Bien Vieillir (BBV) . Ce système d’informations d’un nouveau type a permis la constitution d’une cohorte de santé publique à partir de la population des résidents de la BBV et l’organisation d’un suivi longitudinal syndromique de celle-ci. La BBV a également été évaluée scientifiquement dans un cadre de surveillance et de recherche en santé publique au travers d’une analyse de l’existant : contenu, périodicité, qualité des données. La cohorte construite a ainsi permis la constitution d’un outil de surveillance. Cet échantillon de population a été suivi en temps réel au moyen des fréquences quotidiennes d’apparitions des 26 syndromes des résidents. La méthodologie d’évaluation était celle des systèmes de surveillance sanitaire proposée par le CDC d’Atlanta et a été utilisée pour les syndromes grippaux et les gastro entérites aiguës. Ce fut l’objet du second article.- Enfin la construction d’un nouvel outil de santé publique : la distribution de chacun des syndromes dans le temps (dates de transmissions) et l’espace (les EHPAD de transmissions) a ouvert le champ de la recherche à de nouvelles méthodes d’exploration des données et permis d’étudier plusieurs problématiques liées à la personne âgée : chutes répétées, cancer, vaccinations et fin de vie. / French population is rapidly aging. Senior citizens ratio is increasing and our society needs to rethink its organization, taking into account this change, better knowing this fast growing population group.Even if numerous cohorts of elderly people already exist worldly with four in France and, even as they live in growing numbers in nursing homes and out-patient treatment clinics, knowledge of this population segment is still missing.Today several health and medico-social structures groups as Korian and Orpéa invest in big relational data bases enabling them to get real-time information about their patients/residents. Since 2010 all Korian residents’ files are dematerialized and accessible by requests. They contain at the same time, structured medico-social data describing the residents as well as their treatments and pathologies, but also free-textual data detailing their daily care by the medical staff.Through time and as the computerized resident file (DRI) was mainly conceived as a data base management application, it appeared essential to mine these data and build a decision-making tool intended to improve the care efficiency. The Ageing Well Institute becoming meanwhile the Korian Ageing Well Foundation chose then, working in a private/public partnership, to finance a research work intented to better understand these datas’ informative potential, to assess their reliability and response to public health threats. This research work and this thesis were then designed in several steps:- First, a content analysis of the data warehouse DRI, the objective being to build a research data base, with a social side and a health side. This was the first paper subject.- Then, by direct extraction of the residents’ socio-demographic information at nursing home (NH) entry, adding hospitalizations and deaths, and finally, by an iterative textual extraction process of the transmissions data and by using the Delphi method, we created twenty-four syndromes, added hospitalizations and deaths and built a syndromic data base, the Ageing Well data base. This information system of a new kind, allowed the constitution of a public health cohort for elderly people from the BBV residents’population and its syndromic longitudinal follow-up. The BBV was also scientifically assessed for surveillance and public health research through present situation analysis: content, periodicity and data quality. This cohort then gave us the opportunity to build a surveillance tool and follow the residents’ population in real-time by watching their 26 daily frequency syndromic distributions. The methodology for that assessment, Atlanta CDCs’ health surveillance systems method, was used for flu and acute gastro enteritis syndroms and was the second paper subject.- Finally, the building of a new public health tool: each syndrom’s distribution through time (transmissions dates) and space (transmissions NH ids) opened the research field to new data exploration methods. I used these to study different health problems afflicting senior citizens: frequent falls, cancer, vaccinations and the end of life.

The declining population of black rhinoceroses in Tsavo West national park, Kenya, has served as the driving force behind Project Ngulia, with Ngulia serving as an enclosed area within the park. As of now, the area is equipped with multiple cameras connected to a system that automatically classify animals and humans. This thesis aims to investigate the suitability of the Insta360 One X2 camera acting as a virtual watch tower for capturing and streaming 360° images. This will work in real-time, providing a remote surveillance experience for the park rangers thereby optimizing their work. A system was implemented to create a efficient workflow, which includes stitching of the 360° images, file transfer protocol for image transmission and storage, as well as socket programming to facilitate port monitoring and communication. Additionally, the compat- ibility of two single board computers, LattePanda and Rock 4 SE, with the implemented system was evaluated. User experience methods as field studies, workshops and a user interview were also performed. The work has been developed in Sweden, resulting in limited availability for testing at the target location during the initial months. The outcome was a both locally and remotely working system, together with LattePanda, capturing images of the waterhole in Ngulia. However, because of the conclusions drawn regarding the power supply and the lack of essential functions in the 360° camera, the system was taken home for further research. Propositions is presented regarding future work, some being that the projects within Ngulia team may collaborate to enhance hardware efficiency and explore the utilization of 360° images in educational and entertainment contexts.

O cultivo de camarão branco Litopennaeus vannamei tem provado ser um sector promissor para a economia do nordeste do Brasil. Contudo, a criação de camarão branco no Brasil tem sido afetada negativamente pela ocorrência de doenças virais, ameaçando a sua expansão e sustentabilidade. Por esta razão, depreende-se a importância da elaboração de um sistema de vigilância capaz de detectar e definir a ausência de doenças virais de elevado impacto econômico. O modelo estocástico AquaVigil é aqui implementado para avaliar prospectivamente diferentes estratégias de vigilância para determinar a ausência de doença e identificar a estratégia exigindo menor esforço de amostragem e simultaneamente, fazer o melhor uso dos recursos disponíveis através da implementação de vigilância baseada em risco. O estudo apresentado exemplifica a aplicação regional do sistema proposto para o estado do Ceará, podendo ser aplicado a outros estados do Brasil. O modelo AquaVigil pode analisar qualquer sistema de vigilância baseada em risco semelhante àquele aqui considerado. A criação de camarão no nordeste do Brasil tem sido alvo de vários desafios, desde a ocorrência de doenças virais a mudanças no acesso aos mercados internacionais. Tendo em consideração as dificuldades encontradas pela aquicultura de camarão no nordeste do Brasil, facilmente se compreende a importância de caracterizar e melhor compreender este setor e assim assegurar o seu desenvolvimento sustentável. Para este fim, foram aplicados métodos de análise de correspondência multipla e clustering particional a dados recolhidos durante um levantamento nacional de fazendas de carcinicultura de forma a obter informação necessária para caracterizar tendências e identificar falhas e necessidades existentes. Esta informação será útil no momento de melhorar o manejo das fazendas e elaborar legislação a favor do desenvolvimento do setor / The farming of Pacific white shrimp Litopennaeus vannamei in northeast Brazil, has proven to be a promising sector. However, the farming of Pacific white shrimp in Brazil has been affected negatively by the occurrence of viral diseases, threatening this sector\'s expansion and sustainability. For this reason, the drafting of a surveillance system for early detection and definition of freedom from viral diseases, whose occurrence could result in high economic loses, is of the utmost importance. The stochastic model AquaVigil was implemented to prospectively evaluate different surveillance strategies to determine freedom from disease and identify the strategy with the lowest sampling efforts, making the best use of available resources through risk-based surveillance. The worked example presented was designed for regional application for the state of Ceará and can easily be applied to other Brazilian states. The AquaVigil model can analyze any risk-based surveillance system that considers a similar outline to the strategy here presented. In recent years, shrimp aquaculture has faced many challenges, ranging from the occurrence of viral diseases to changes in market access. Considering the past and present challenges faced by the shrimp farmers in Northeast Brazil it is easily understood that the comprehensive characterization of the shrimp farming is of the utmost importance when striving for sustainable development. To this aim, the exploratory data analysis methods of multiple correspondence analysis and partitional clustering were applied to the data collected through a national census to extract the greatest amount of information and profile shrimp farms, identifying gaps and needs. The results of the analysis will contribute to improve management practices and policy-making for sustainable shrimp farming in Northeast Brazil

Introduktion: Vårdrelaterade infektioner (VRI) har negativ inverkan på folkhälsan med påverkan på mortalitet, morbiditet och livskvalitet. Inom hälso- och sjukvården är VRI en stor utmaning. Den vanligaste VRI i Sverige är vårdrelaterade urinvägsinfektioner (VUVI). För att förebygga VRI är regelbundna mätningar och återkopplingar viktigt. Mätinstrument som idag används på lokal och nationell nivå för att mäta incidens av VRI är bl.a. Markörbaserad journalgranskning och Infektionsverktyget. Syfte: Validera Infektionsverktyget och jämföra incidens av VUVI från Infektionsverktyget mot manuell journalgranskning och Markörbaserad journalgranskning samt studera förekomsten av urinkateter i samband med VUVI. Metod: Under 2017 samlades data in från 143 slumpmässigt utvalda journaler i slutenvården på Södra Älvsborgs sjukhus. Diagnostiserad VUVI och samhällsförvärvad urinvägsinfektion (SUVI) i Infektionsverktyget validerades mot manuell journalgranskning. Incidensen av VUVI i Infektionsverktyget jämfördes även mot VUVI i markörbaserad journalgranskning. Resultat: Incidens av VUVI innan validering var 1.5 % medan den uppskattade incidensen efter manuell journalgranskning och validering var 3.6 %. Markörbaserad journalgranskning visade en incidens på 1.1 %. I 65.6 % av fallen med VUVI fanns en koppling till urinkateter. Den mest förekommande orsaken till inkorrekt registreringen av VUVI var att patienter med urinkateter bedömdes som SUVI. Slutsats: Den rapporterade incidensen av VUVI skiljer sig mellan de mätinstrument som används idag och incidensen är troligtvis högre än vad som idag rapporteras. För att få bra kvalitet på övervakningsdata krävs kunskap och granskning av hög kvalitet. Dock begränsas resultatet från denna studie av att studiepopulationen var relativt liten och därmed begränsar generaliserbarheten. / Introduction: Health care associated infections (HCAIs) have a negative impact on public health, with an impact on mortality, morbidity and quality of life. The most common HCAIs in Sweden are health care associated urinary tract infections (UTI). One important component in preventing HCAIs are regular measurements and feedback to these. Instruments that are currently used for measuring HCAIs at a local and national level in Sweden are marker-based journal review and the Anti-Infection Tool (AIT). Aim:  The aim of this study was to validate the AIT and compare the incidence of health care associated UTI measured with the AIT against both manual journal review and marker-based journal review. The aim was also to study the presence of urinary catheters in conjunction with a healthcare associated UTI. Methods: In 2017, data was collected from 143 records from a random sample of patients admitted to somatic wards at Södra Älvsborg's Hospital. From the AIT, diagnosed health care associated UTI and community-acquired UTI were studied and validated against manual journal review. Also, the incidence of health care associated UTI in the AIT was compared with that found in marker-based journal review. Results: Incidence of health care associated UTI before the validation was 1.5% while the estimated incidence after manual journal review and validation was 3.6%. Marker-based journal review showed an incidence of 1.1%. In 65.6 % of the cases with health care associated UTI, the patient was equipped with urinary catheter. The most common cause of incorrect registration of health care associated UTI was that patients with urinary catheters were assessed as community-acquired UTI.   Conclusion:  The reported incidence of health care associated UTI differs greatly between these instruments. The incidence is probably much higher than what is currently reported using these instruments. To obtain good quality of monitoring data, knowledge and journal reviews of high quality are required. However, the generalizability of the result of this study is limited, due to the relatively small study sample.

O cultivo de camarão branco Litopennaeus vannamei tem provado ser um sector promissor para a economia do nordeste do Brasil. Contudo, a criação de camarão branco no Brasil tem sido afetada negativamente pela ocorrência de doenças virais, ameaçando a sua expansão e sustentabilidade. Por esta razão, depreende-se a importância da elaboração de um sistema de vigilância capaz de detectar e definir a ausência de doenças virais de elevado impacto econômico. O modelo estocástico AquaVigil é aqui implementado para avaliar prospectivamente diferentes estratégias de vigilância para determinar a ausência de doença e identificar a estratégia exigindo menor esforço de amostragem e simultaneamente, fazer o melhor uso dos recursos disponíveis através da implementação de vigilância baseada em risco. O estudo apresentado exemplifica a aplicação regional do sistema proposto para o estado do Ceará, podendo ser aplicado a outros estados do Brasil. O modelo AquaVigil pode analisar qualquer sistema de vigilância baseada em risco semelhante àquele aqui considerado. A criação de camarão no nordeste do Brasil tem sido alvo de vários desafios, desde a ocorrência de doenças virais a mudanças no acesso aos mercados internacionais. Tendo em consideração as dificuldades encontradas pela aquicultura de camarão no nordeste do Brasil, facilmente se compreende a importância de caracterizar e melhor compreender este setor e assim assegurar o seu desenvolvimento sustentável. Para este fim, foram aplicados métodos de análise de correspondência multipla e clustering particional a dados recolhidos durante um levantamento nacional de fazendas de carcinicultura de forma a obter informação necessária para caracterizar tendências e identificar falhas e necessidades existentes. Esta informação será útil no momento de melhorar o manejo das fazendas e elaborar legislação a favor do desenvolvimento do setor / The farming of Pacific white shrimp Litopennaeus vannamei in northeast Brazil, has proven to be a promising sector. However, the farming of Pacific white shrimp in Brazil has been affected negatively by the occurrence of viral diseases, threatening this sector\'s expansion and sustainability. For this reason, the drafting of a surveillance system for early detection and definition of freedom from viral diseases, whose occurrence could result in high economic loses, is of the utmost importance. The stochastic model AquaVigil was implemented to prospectively evaluate different surveillance strategies to determine freedom from disease and identify the strategy with the lowest sampling efforts, making the best use of available resources through risk-based surveillance. The worked example presented was designed for regional application for the state of Ceará and can easily be applied to other Brazilian states. The AquaVigil model can analyze any risk-based surveillance system that considers a similar outline to the strategy here presented. In recent years, shrimp aquaculture has faced many challenges, ranging from the occurrence of viral diseases to changes in market access. Considering the past and present challenges faced by the shrimp farmers in Northeast Brazil it is easily understood that the comprehensive characterization of the shrimp farming is of the utmost importance when striving for sustainable development. To this aim, the exploratory data analysis methods of multiple correspondence analysis and partitional clustering were applied to the data collected through a national census to extract the greatest amount of information and profile shrimp farms, identifying gaps and needs. The results of the analysis will contribute to improve management practices and policy-making for sustainable shrimp farming in Northeast Brazil

This thesis addresses a multi-agent search problem where several agents, equipped with sensors and communication devices, search an unknown area. Lack of information about the search space is modeled as an uncertainty density distribution. A sequential deploy and search (SDS) strategy is formulated where the agents are first deployed to maximize single step search effectiveness. To achieve an optimal deployment, a multi-center objective function defined using the Voronoi cells and the uncertainty distribution is optimized. It is shown that the critical points of this objective function are the centroids of the Voronoi cells. A proportional control law is proposed that makes the agents move to their respective “centroids”. Assuming agents to be first order dynamical systems and using LaSalle's invariance principle, it is shown that the closed-loop system converges globally asymptotically to the critical points. It is also shown that the sequential deploy and search strategy is spatially distributed with respect to the Delaunay graph corresponding to any given agent configuration. Next, a combined deploy and search (CDS) strategy is proposed where, instead of first deploying agents and then performing the search, the agents engage in search operation as they move toward the centroids. This strategy gives rise to shorter agent trajectories compared to the SDS strategy. Then the problem is formulated with practical constraints such as sensor range limits and limit on maximum speed of the agents. A few issues relating to implementation of the proposed search strategies are also addressed. Finally, the assumption of homogeneous agents is relaxed and agents equipped with sensors with heterogeneous capabilities are considered. A generalized Voronoi partitioning scheme is proposed and used to formulate a heterogeneous locational optimization problem. In this problem the agents are deployed in the search space optimizing the sensor effectiveness. As earlier, the two search strategies are proposed. Simulation experiments are carried out to validate the performance of the proposed search strategies. The simulation results indicate that both the proposed search strategies perform quite well even when the conditions deviated from the nominal. It is also shown that the combined deploy and search strategy leads to shorter and smoother trajectories than those of the sequential deploy and search strategy with the same parameters.