This thesis considers two related problems that has application in the field of experimental design for clinical trials: • fixed sample size determination for parallel arm, double-blind survival data analysis to test the hypothesis of no difference in survival functions, and • blinded sample size re-estimation for the same. For the first problem of fixed sample size determination, a method is developed generally for testing of hypothesis, then applied particularly to survival analysis; for the second problem of blinded sample size re-estimation, a method is developed specifically for survival analysis. In both problems, the exponential survival model is assumed. The approach we propose for sample size determination is Bayesian decision theoretical, using explicitly a loss function and a prior distribution. The loss function used is the intrinsic discrepancy loss function introduced by Bernardo and Rueda (2002), and further expounded upon in Bernardo (2011). We use a conjugate prior, and investigate the sensitivity of the calculated sample sizes to specification of the hyper-parameters. For the second problem of blinded sample size re-estimation, we use prior predictive distributions to facilitate calculation of the interim test statistic in a blinded manner while controlling the Type I error. The determination of the test statistic in a blinded manner continues to be nettling problem for researchers. The first problem is typical of traditional experimental designs, while the second problem extends into the realm of adaptive designs. To the best of our knowledge, the approaches we suggest for both problems have never been done hitherto, and extend the current research on both topics. The advantages of our approach, as far as we see it, are unity and coherence of statistical procedures, systematic and methodical incorporation of prior knowledge, and ease of calculation and interpretation. / Statistics

Detecting and tracking moving objects are important topics in computer vision research. Classical methods perform well in applications of steady cameras. However, these techniques are not suitable for the applications of moving cameras because the unconstrained nature of realistic environments and sudden camera movement makes cues to object positions rather fickle. A major difficulty is that every pixel moves and new background keeps showing up when a handheld or car-mounted camera moves. In this dissertation, a novel estimation method of camera motion parameters will be discussed first. Based on the estimated camera motion parameters, two detection algorithms are developed using Bayes' rule and belief propagation. Next, an MCMC-based feature-guided particle filtering method is presented to track detected moving objects. In addition, two detection algorithms without using camera motion parameters will be further discussed. These two approaches require no pre-defined class or model to be trained in advance. The experiment results will demonstrate robust detecting and tracking performance in object sizes and positions.

In recent years network intrusions have become a severe threat to the privacy and safety of computer users. Recent cyber attacks compromise a large number of hosts to form botnets. Hackers not only aim at harvesting private data and identity information from compromised nodes, but also use the compromised nodes to launch attacks such as distributed denial-of-service (DDoS) attacks. As a counter measure, Intrusion Detection Systems (IDS) are used to identify intrusions by comparing observable behavior against suspicious patterns. Traditional IDSs monitor computer activities on a single host or network traffic in a sub-network. They do not have a global view of intrusions and are not effective in detecting fast spreading attacks, unknown, or new threats. In turn, they can achieve better detection accuracy through collaboration. An Intrusion Detection Network (IDN) is such a collaboration network allowing IDSs to exchange information with each other and to benefit from the collective knowledge and experience shared by others. IDNs enhance the overall accuracy of intrusion assessment as well as the ability to detect new intrusion types. Building an effective IDN is however a challenging task. For example, adversaries may compromise some IDSs in the network and then leverage the compromised nodes to send false information, or even attack others in the network, which can compromise the efficiency of the IDN. It is, therefore, important for an IDN to detect and isolate malicious insiders. Another challenge is how to make efficient intrusion detection assessment based on the collective diagnosis from other IDSs. Appropriate selection of collaborators and incentive-compatible resource management in support of IDSs' interaction with others are also key challenges in IDN design. To achieve efficiency, robustness, and scalability, we propose an IDN architecture and especially focus on the design of four of its essential components, namely, trust management, acquaintance management, resource management, and feedback aggregation. We evaluate our proposals and compare them with prominent ones in the literature and show their superiority using several metrics, including efficiency, robustness, scalability, incentive-compatibility, and fairness. Our IDN design provides guidelines for the deployment of a secure and scalable IDN where effective collaboration can be established between IDSs.

In recent years network intrusions have become a severe threat to the privacy and safety of computer users. Recent cyber attacks compromise a large number of hosts to form botnets. Hackers not only aim at harvesting private data and identity information from compromised nodes, but also use the compromised nodes to launch attacks such as distributed denial-of-service (DDoS) attacks. As a counter measure, Intrusion Detection Systems (IDS) are used to identify intrusions by comparing observable behavior against suspicious patterns. Traditional IDSs monitor computer activities on a single host or network traffic in a sub-network. They do not have a global view of intrusions and are not effective in detecting fast spreading attacks, unknown, or new threats. In turn, they can achieve better detection accuracy through collaboration. An Intrusion Detection Network (IDN) is such a collaboration network allowing IDSs to exchange information with each other and to benefit from the collective knowledge and experience shared by others. IDNs enhance the overall accuracy of intrusion assessment as well as the ability to detect new intrusion types. Building an effective IDN is however a challenging task. For example, adversaries may compromise some IDSs in the network and then leverage the compromised nodes to send false information, or even attack others in the network, which can compromise the efficiency of the IDN. It is, therefore, important for an IDN to detect and isolate malicious insiders. Another challenge is how to make efficient intrusion detection assessment based on the collective diagnosis from other IDSs. Appropriate selection of collaborators and incentive-compatible resource management in support of IDSs' interaction with others are also key challenges in IDN design. To achieve efficiency, robustness, and scalability, we propose an IDN architecture and especially focus on the design of four of its essential components, namely, trust management, acquaintance management, resource management, and feedback aggregation. We evaluate our proposals and compare them with prominent ones in the literature and show their superiority using several metrics, including efficiency, robustness, scalability, incentive-compatibility, and fairness. Our IDN design provides guidelines for the deployment of a secure and scalable IDN where effective collaboration can be established between IDSs.

Stora och Lilla Ullfjärden är två sjöar som tillhör den innersta delen av en vik i Mälaren. Här har den höga fosforbelastningen länge varit ett problem då den leder till kraftig algblomning varje år. Hittills har åtgärder mot fosforbelastningen riktats mot utsläppskällor av fosfor från omgivande mark, men dessa åtgärder har inte varit tillräckliga. Det beror på att fosfor har lagrats på sjöarnas bottnar under många år vilket orsakar intern fosforbelastning.  Restaureringsbehovet i sjöarna bestämdes genom att hitta den totalfosforhalt som krävs för att uppnå god ekologisk vattenstatus. Det är önskvärt att sjöarna uppnår god status med avseende på klorofyll, siktdjup och växtplankton. Dessa parametrar uppvisar starka samband till halten totalfosfor och på så sätt kunde målbilden för halten totalfosfor bestämmas till 9,4 µg/l i Stora Ullfjärden och 8,7 µg/l i Lilla Ullfjärden, med siktdjupet som den begränsande parametern.  En multikriterieanalys togs fram där det undersöktes vilken metod som är den lämpligaste sjörestaureringsåtgärden mot internbelastningen. Kostnader, effektivitet och livslängd för sex välstuderade åtgärdsmetoder fastställdes genom en metaanalys. Resultatet visar att aluminiumbehandling verkar vara det lämpligaste valet av åtgärdsmetod. Åtgärden innebär att en aluminiumlösning tillsätts sjöarnas sediment vilket binder fosforn och gör så att den inte frigörs.  Det undersöktes även om kostnaden för aluminiumbehandlingen kan motiveras genom att belysa sjöarnas ekonomiska värden. Under arbetet togs det fram olika typer av underlag för att bedöma sjöarnas värde. Genom att lista de tjänster som sjöarna bidrar med och tilldela tjänsterna ett uppskattat ekonomiskt värde konstaterades att sjöarna framför allt har ett högt naturvärde och bidrar med många kulturella ekosystemtjänster såsom bad, fiske och rekreation. Utifrån en tidigare studie uppskattades allmänhetens betalningsvilja för att uppnå god vattenstatus i Stora och Lilla Ullfjärden till 23 miljoner kronor. Det betyder att allmänhetens betalningsvilja täcker kostnaderna för aluminiumbehandlingen i Stora och Lilla Ullfjärden, vilket uppskattades till 22 miljoner kronor.  Med hjälp av bayesiansk beslutsanalys har det även kunnat konstateras att sjöarna bör aluminiumbehandlas om det minskade värdet för att god status inte uppnås (såsom minskat estetiskt värde, minskad biologisk mångfald, försämrad förmåga att rena vatten, med mera) anses kosta mer än 25 miljoner kronor på en 12 årsperiod. Eftersom beslutsmaterialet har en del osäkerheter finns det ett fortsatt behov av ytterligare undersökningar. / Stora Ullfjärden and Lilla Ullfjärden are two lakes that belong to the innermost part of a bay in Lake Mälaren. Phosphorus load has been a problem for a long time as it causes large algal bloom every summer. Attempt to decrease the phosphorus load in the lakes have so far been focused on measures to control the phosphorous emissions from the surrounding land. However, these measurers have not been sufficient. This is because phosphorus has been accumulated in the bottom sediment of the lakes for many years which cause internal phosphorus loading.  The amount of phosphorous needed to be removed was determined by identifying the maximum total phosphorus content the lake can handle while still demonstrating good ecological status. It is desirable that the lakes achieve good status regarding chlorophyll, secchi depth and phytoplankton. These three parameters show strong correlations to the total phosphorus content and in this way the desired future total phosphorus content could be calculated to 9,4 µg/l in Stora Ullfjärden and 8,7 µg/l in Lilla Ullfjärden, with the secchi depth as the limiting parameter. A multi-criteria decision analysis was developed to determine the most suitable lake restoration measure to reduce internal phosphorous loads. Costs, efficacy, and longevity of six well-known mitigation measures were determined by a meta-analysis. The results verify that aluminum treatment appears to be the most suitable choice. The method involves adding aluminum solution to the lakes' sediment, which binds phosphorus, so it is not released.  Additionally, it was examined whether the cost of aluminum treatment can be justified by highlighting the economic values ​​of the lakes. Different types of decision material are presented in the report to assess a value to the lakes. By listing services of the lakes and assign them an estimated economic value, it was determined that the lakes have a high nature value and contribute with many cultural ecosystem services such as swimming, fishing, and recreation. Based on a previous study, the public's willingness to pay to achieve good water status in Stora Ullfjärden and Lilla Ullfjärden was estimated to SEK 23 million. Hence, the public's willingness to pay covers the costs of the aluminum treatment in Stora and Lilla Ullfjärden, which was estimated to SEK 22 million.  Lastly, a Bayesian model for decision making was used. It was discovered that the lakes are recommended to undergo an aluminum treatment if the reduced value for not achieving good status in the lakes (e.g., decreased aesthetic value, reduced biodiversity, and impaired ability to purify water) is considered to cost more than SEK 25 million over a 12-year period. As the decision material has some uncertainties, there is a continuing need for further investigations.

A 2009 report published by the National Research Council addressed the need for improvements in the field of forensic science. In the report emphasis was placed on the need for more rigorous scientific analysis within many forensic science disciplines and for established limitations and determination of error rates from statistical analysis. This research focused on multivariate statistical techniques for the analysis of spectral data obtained for multiple forensic applications which include samples from: automobile float glasses and paints, bones, metal transfers, ignitable liquids and fire debris, and organic compounds including explosives. The statistical techniques were used for two types of data analysis: classification and discrimination. Statistical methods including linear discriminant analysis and a novel soft classification method were used to provide classification of forensic samples based on a compiled library. The novel soft classification method combined three statistical steps: Principal Component Analysis (PCA), Target Factor Analysis (TFA), and Bayesian Decision Theory (BDT) to provide classification based on posterior probabilities of class membership. The posterior probabilities provide a statistical probability of classification which can aid a forensic analyst in reaching a conclusion. The second analytical approach applied nonparametric methods to provide the means for discrimination between samples. Nonparametric methods are performed as hypothesis test and do not assume normal distribution of the analytical figures of merit. The nonparametric iv permutation test was applied to forensic applications to determine the similarity between two samples and provide discrimination rates. Both the classification method and discrimination method were applied to data acquired from multiple instrumental methods. The instrumental methods included: Laser Induced-Breakdown Spectroscopy (LIBS), Fourier Transform Infrared Spectroscopy (FTIR), Raman spectroscopy, and Gas Chromatography-Mass Spectrometry (GC-MS). Some of these instrumental methods are currently applied to forensic applications, such as GC-MS for the analysis of ignitable liquid and fire debris samples; while others provide new instrumental methods to areas within forensic science which currently lack instrumental analysis techniques, such as LIBS for the analysis of metal transfers. The combination of the instrumental techniques and multivariate statistical techniques is investigated in new approaches to forensic applications in this research to assist in improving the field of forensic science.

Orthogonal frequency division multiplexing (OFDM) modulator with redundancy has been adopted in many wireless communication systems for higher data rate transmissions. The introduced redundancy at the transmitter allows us to overcome serious inter-block interference (IBI) problems due to highly dispersive channel. However, the selection of redundancy length will affect the system performance and spectral efficiency, and is highly dependent on the length of channel impulse response. In this thesis, based on the pseudorandom postfix (PRP) OFDM scheme we propose a novel block-based OFDM transceiver framework. Since in the PRP-OFDM system the PRP can be employed for semi-blind channel estimation with order-one statistics of the received signal. Hence, for sufficient redundancy case the PRP-OFDM system with the Bayesian decision feedback equalizer (DFE) is adopted for suppressing the IBI and ISI simultaneously. However, for the insufficient redundancy case (the length of redundancy is less than the order of channel), we first propose a modified scheme for channel estimation. To further reduce the complexity of receiver, the maximum shortening signal-to-noise-ratio time domain equalizer (MSSNR TEQ) with the Bayesian DFE is developed for suppressing the IBI and ISI, separately. That is, after knowing the channel state information (CSI) and removing the effect of IBI with MSSNR TEQ, the Bayesian DFE is applied for eliminating the ISI. Via computer simulation, we verify that performance improvement, in terms of bit error rate (BER), compared with the conventional block-based minimum mean square error (MMSE)-DFE can be achieved.

The ability to solve problems creatively has been crucial for the adaptation and survival of humans throughout history. In many real–life situations, cognitive processes are not isolated. Humans are social, they communicate and form groups to solve daily problems and make decisions. Therefore, the final output of cognitive processes can come from multi–brains in groups rather than an individual one. This multi–brain output can be largely different from the output that an individual person produces in isolation. As a result, it is essential to include team–level processes in cognitive models to make a more accurate description of real– world cognitive processes in general and problem solving in particular. This research aims to answer the general question of how working in a team affects creative problem solving. For doing that, first, we propose a computational model for multi-agent creative problem solving. Then, we show how the model can be used to study the factors that are involved in creativity in teams and potentially will suggest answers to questions such as, ‘how team size is related to creativity’.

From the Proceedings of the 1975 Meetings of the Arizona Section - American Water Resources Assn. and the Hydrology Section - Arizona Academy of Science - April 11-12, 1975, Tempe, Arizona / Bayesian decision theory is a method for comparing expected utilities of alternative actions given various possible states of nature. The method treats uncertainty as to the true state of nature by determining the expected utility of each action in terms of the probabilities of the various possible states. The decision rule is to choose the action having the best expected utility. This paper illustrates an application of Bayesian decision theory in a well field design problem where a decision had to be made regarding capacity-density combination for wells located in an extensive uniform grid. The uncertainty lay in anticipating the frequencies of transmissivity values among the wells.

The reported research provides a rough guide to the best practice of decision modeling concerning urban pipeline systems’ rehabilitation. The thesis aims to bring attention to the fact that a proper decision-making model is a cornerstone for efficient infrastructure management. More precisely, this thesis aims to increase the knowledge about applicable decision support methods by identifying relevant factors that should be considered in the decision-making process. This can, facilitate future rehabilitation attempts of existing urban infrastructure. A utility-based decision model was adopted for a water distribution network in Sweden to locate and rehabilitate leakages as an ultimate sign of failure. This was performed by implementing and evaluating a Bayesian decision model including the treatment of uncertainties in evaluating the best decision from a short-term perspective. Despite its simplicity, the result showed that the proposed model could facilitate problem-solving approaches when uncertainty is an issue. Considering the several interacting factors of services and the availability of information, the importance of problem structuring before applying a decision model was extensively acknowledged. As a result, a conceptual decision model was proposed to choose the most appropriate decision model applicable for a particular problem in the essence of deciding how to decide. The presented model illustrated the first steps of developing a theoretical framework for a rational yet practical decisionmaking. This approach, which is aimed to be further employed in rehabilitation strategies of urban pipelines, ensures that the chosen decision technique has explicitly considered different levels of uncertainty and would be the best-established solution for a particular type of problem, organization, and stakeholder. This effort may help the decision analysts define the problem and elicit objectives and values relatively early in the decision-making to ensure that decisions to be selected would support the desired outcomes, actions, and core values. Then, a critical evaluation of the decision strategy was presented by comparing the performed Bayesian approach with the proposed conceptual model. Then so, it was shown that the choice of the decision model is dissimilar if the presented specific basic components vary. This was performed by presenting two semi-fictitious case studies, exemplifying the framework’s importance in structuring the assessment of available means. / Forskningen som redovisas i denna uppsats utgör en översiktlig guide till en praktisktillämpning av beslutsmodellering gällande underhåll av urbana ledningssystem.Syftet med licentiatuppsatsen är att betona att en korrekt modell för beslutsfat-tande är nödvändig för en effektiv förvaltning av infrastruktur. Mer specifikt ärmålet att öka kunskapen om tillämpbara beslutsstödsmetoder genom att identifiera relevanta faktorer som bör beaktas i beslutsprocessen. Det förväntas underlätta framtida underhållsaktiveter för befintlig urban infrastruktur. En nyttobaserad beslutsmodell för åtgärdsplanering har applicerats på en del av ettsvenskt vattenledningssystem, där läckage är den kritiska händelse som hanteras.Modellen baserad på Bayesiansk beslutsteori har implementerats och utvärderatsmed avseende på hantering av osäkerheter och beslutsoptimering ur ett korttidsper-spektiv. Trots modellens enkelhet visar resultatet att den kan underlätta metodvalför problemlösning när det råder osäkerheter i förutsättningarna. Vikten av en tydlig och strukturerad problembeskrivning inför tillämpningen av enbeslutsmodell bekräftas, där beaktande av interaktioner mellan ibland flera faktoreri systemets funktion och den tillgängliga informationen är viktig. Som ett resultatföreslås en konceptuell metod för att välja den mest lämpliga beslutsmodellen förett specifikt problem med syftet att besluta hur man bör besluta. Den presenter-ade metoden utgör ett första steg i utvecklingen av ett teoretiskt ramverk för ettrationellt och samtidigt praktiskt beslutsfattande. Arbetet hjälper beslutsfattarenatt strukturera problemet och lyfta syftet och värden tidigt i beslutsfattandet föratt säkerställa att tagna beslut stödjer eftersökta utfall, åtgärder och kärnvärden. Vidare har en kritisk utvärdering av beslutsstrategier presenterats som en jämförelsemellan den Bayesianska beslutsmodellen och den konceptuella metoden. Den visaratt valet av beslutsmodell skiljer sig om de grundläggande förutsättningarna ärolika. Utvärderingen baseras på två semifiktiva fallstudier som visar på vikten avstrukturering i bedömningen av tillgänglig information och tillgängliga resurser. / <p>2022-10-24</p> / Mistra InfraMaint