Numerous methods to assess student knowledge are present throughout every step of a students€™ education. Skill-based assessments include homework, quizzes and tests while curriculum exams comprise of the SAT and GRE. The latter assessments provide an indication as to how well a student has retained a learned national curriculum however they are unable to identify how well a student performs at a fine grain skill level. The former assessments hone in on a specific skill or set of skills, however, they require an excessive amount of time to collect curriculum-wide data. We've developed a system that assesses students at a fine grain level in order to identify non- mastered skills within each student€™s zone of proximal development. €œPLACEments€� is a graph-driven computer adaptive test which not only provides thorough student feedback to educators but also delivers a personalized remediation plan to each student based on his or her identified non-mastered skills. As opposed to predicting state test scores, PLACEments objective is to personalize learning for students and encourage teachers to employ formative assessment techniques in the classroom. We have conducted a randomized controlled study to evaluate the learning value PLACEments provides in comparison to traditional methods of targeted skill mastery and retention.

I dagens konsumentsamhälle växer hela tiden utbudet av varor och tjänster att välja mellan och därmed växer även konkurrensen mellan de olika företag som finns ute på marknaden.Utvecklingen inom direktmarknadsföringen går snabbt framåt och med framstegen inom området kommer det även fram nya sätt att nå kunderna. Individanpassade erbjudande är ett relativt nytt sätt för företagen att arbeta med i sin marknadsföring vilket gör det extra viktigt för företagen att kontrollera hur deras erbjudanden tas emot av deras kunder.Det finns många fördelar med att använda sig av personanpassad DM men självkart kan det också uppstå problem vid användandet av den här typen av erbjudanden. Ett av de främsta problemen är att det kan inkräkta den personliga integriteten hos den som mottar dem.Vår studie är en fallstudie där vi valt att inrikta oss på ICA och hur de arbetar med sin DM. Vi har valt att göra en kundundersökning med respondenter som har ICA-kort, där vi ställer frågor kring kundernas uppfattning av företagets lagring av köpbeteende och anpassade erbjudanden. Genom att göra en intervju med vår kontaktperson på företaget som har en ledande position på ICA:s marknadsavdelning, har vi fått information som hjälpt oss besvara en del av vår problemfrågeställning.Vårt syfte med studien är att undersöka och beskriva hur ICA arbetar med sin direktmarknadsföring. Vi vill skapa en förståelse kring ICA:s förhoppningar om hur deras direktmarknadsföring uppfattas av kunderna, samt om det stämmer överens med verkligheten.De problemformuleringar vi arbetat efter ser ut på följande vis:Huvudfråga: Hur uppfattas ICA:s DM-erbjudande ”mina varor” av deras kunder?Delfråga 1: Hur arbetar ICA med sin DM?Delfråga 2: Hur vill och tror ICA att deras DM-erbjudande ”mina varor” uppfattas av deras kunder?Vår undersökning visar att ICA:s personanpassade erbjudande ”mina varor” överlag är en uppskattad tjänst bland deras kunder. Problem som kan uppstå gällande inkräktande av den personliga integriteten var inte så stora som företaget själva trodde att det kunde bli vid uppstartandet av den här typen av erbjudande. Majoriteten av respondenterna i vår undersökning har angivit att de var positiva till ”mina varor”. Den största andelen kunder uppfattade också de här erbjudandena som något de ibland har nytta av och även ibland att de är relevanta för dem och deras hushåll.

Då konkurrensen mellan företag inom kläd- och modebranschen är hög erbjuder mångaföretag kundklubbar och lojalitetsprogram till sina kunder. Kundklubbar ska vara ett sättför företag att differentiera sig, men blir det en differentiering när de flesta företagerbjuder sina kunder liknande kundklubbar och lojalitetsprogram?För att differentiera sig har företag inom branschen utvidgat sina kundklubbar ochlojalitetsprogram, men med åtgärder som de flesta företag kan använda inom branschen,vilket gör att differentieringen uteblir. Den svenska matvarukedjan ICA har tagitkundklubben till en ny nivå och erbjuder sina kunder ”mina varor”. Utifrån kunderstidigare köp får de sedan erbjudanden som är personliga. Varje erbjudande blir unikt.Frågan är om det finns ett intresse hos konsumenter inom kläd- och modebranschen förett anpassat erbjudande, för att göra kundklubbars erbjudanden mer unika. För att det skagenomföras behöver företag registrera kunders köpvanor och beteenden. Kan erbjudandetdå utgöra ett hot mot kunders integritet?Syftet med uppsatsen är att undersöka vad kunder inom kläd- och modebranschen anserom dagens kundklubbar och hur de ställer sig till att införa personanpassade erbjudanden,likt ICA:s, i kläd- och modebranschen. Målet är att undersöka om kunden har ett intresseför en ny form av kundklubb i kläd- och modebranschen.Frågeställningarna som ska undersökas är:Huvudfråga:Hur ser kunden på personanpassade erbjudanden i kläd- och modebranschen?Underfrågor:Hur ser kunden på de erbjudanden som kundklubbar erbjuder idag?Hur påverkas kunder av det faktum att företag måste registrera deras köpvanor för attkunna erbjuda personliga erbjudanden?För att besvara frågeställningarna har en fallstudie utförts med hjälp av en fokusgruppsamt fem personliga intervjuer. Undersökningen har visat att konsumenter inom kläd- ochmodebranschen relaterar kundklubbar med erbjudanden för att få dig som kund att handlamer, vilket konsumenterna i vissa fall också gör. Det visade sig också att konsumenterinom branschen inte tror att det skulle vara lika lätt att anpassa personliga erbjudanden påkläder och mode som det är på dagligvaror. Dessutom är dagens konsumenter vana vidatt företag registrerar deras köp och tycker att det är acceptabelt så länge företag intesamlar för personlig information i marknadsföringssyfte.

Personalized cancer medicine is a cancer care paradigm in which diagnostic and therapeutic strategies are customized for individual patients. Microsystems that are created by Micro-Electro-Mechanical Systems (MEMS) technology and integrate various diagnostic and therapeutic methods on a single chip hold great potential to enable personalized cancer medicine. Toward ultimate realization of such microsystems, this thesis focuses on developing critical functional building blocks that perform genetic variation identification (single-nucleotide polymorphism (SNP) genotyping) and specific, efficient and flexible cell manipulation on microfluidic surfaces. For the identification of genetic variations, we first present a bead-based approach to detect single-base mutations by performing single-base extension (SBE) of SNP specific primers on solid surfaces. Successful genotyping of the SNP on exon 1 of HBB gene demonstrates the potential of the device for simple, rapid, and accurate detection of SNPs. In addition, a multi-step solution-based approach, which integrates SBE with mass-tagged dideoxynucleotides and solid-phase purification of extension products, is also presented. Rapid, accurate and simultaneous detection of 4 loci on a synthetic template demonstrates the capability of multiplex genotyping with reduced consumption of samples and reagents. For cell manipulation, we first present a microfluidic device for cell purification with surface-immobilized aptamers, exploiting the strong temperature dependence of the affinity binding between aptamers and cells. Further, we demonstrate the feasibility of using aptamers to specifically separate target cells from a heterogeneous solution and employing environmental changes to retrieve purified cells. Moreover, spatially specific capture and selective temperature-mediated release of cells on design-specified areas is presented, which demonstrates the ability to establish cell arrays on pre-defined regions and to collect only specifically selected cell groups for downstream analysis. We also investigate tunable microfluidic trapping of cells by exploiting the large compliance of elastomers to create an array of cell-trapping microstructures, whose dimensions can be mechanically modulated by inducing uniform strain via the application of external force. Cell trapping under different strain modulations has been studied, and capture of a predetermined number of cells, from single cells to multiple cells, has been achieved. In addition, to address the lack of aptamers for targets of interest, which is a major hindrance to aptamer-based cell manipulation, we present a microfluidic device for synthetically isolating cell-targeting aptamers from a randomized single-strand DNA (ssDNA) library, integrating cell culturing with affinity selection and amplification of cell-binding ssDNA. Multi-round aptamer isolation on a single chip has also been realized by using pressure-driven flow. Finally, some perspectives on future work are presented, and strategies and notable issues are discussed for further development of MEMS/microfluidics-based devices for personalized cancer medicine.

This work takes us a step closer to realizing personalized medicine, complementing empirical and heuristic way in which clinicians typically work. This thesis presents mechanistic models of physiology. These models, given continuous signals from a patient, can be fine-tuned via parameter estimation methods so that the model's outputs match the patient's. We thus obtain a virtual patient mimicking the patient at hand. Therapeutic scenarios can then be applied and optimal diagnosis and therapy can thus be attained. As such, personalized medicine can then be achieved without resorting to costly genetics. In particular we have developed a novel comprehensive mathematical model of the cardiopulmonary system that includes cardiovascular circulation, respiratory mechanics, tissue and alveolar gas exchange, as well as short-term neural control. Validity of the model was proven by the excellent agreement with real patient data, under normo-physiological as well as hypercapnic and hypoxic conditions, taken from literature. As a concrete example, a submodel of the lung mechanics was fine-tuned using real patient data and personalized respiratory parameters (resistance, R_rs, and compliance, C_rs) were estimated continually. This allows us to compute the patient's effort (Work of Breathing), continuously and more importantly noninvasively. Finally, the use of Bayesian estimation techniques, which allow incorporation of population studies and prior information about model's parameters, was proposed in the contest of patient-specific physiological models. A Bayesian Maximum a Posteriori Probability (MAP) estimator was implemented and applied to a case-study of respiratory mechanics. Its superiority against the classical Least Squares method was proven in data-poor conditions using both simulated and real animal data. This thesis can serve as a platform for a plethora of applications for cardiopulmonary personalized medicine.

Mobile devices, such as smartphones and wearable devices, have become a popular platform to deliver recommendations and interact with users. To learn the decision rule of assigning recommendations, i.e. policy, neither one homogeneous policy for all users nor completely heterogeneous policy for each user is appropriate. Many attempts have been made to learn a policy for making recommendations using observational mobile health (mHealth) data. The majority of them focuses on a homogeneous policy, that is a one-fit-to-all policy for all users. It is a fair starting point for mHealth study, but it ignores the underlying user heterogeneity. Users with similar behavior pattern may have unobservable underlying heterogeneity. To solve this problem, we develop a personalized learning framework that models both population and personalized effect simultaneously. In the first part of this dissertation, we address the personalized policy learning problem using longitudinal mHealth application usage data. Personalized policy represents a paradigm shift from developing a single policy that may prescribe personalized decisions by tailoring. Specifically, we aim to develop the best policy, one per user, based on estimating random effects under generalized linear mixed model. With many random effects, we consider new estimation method and penalized objective to circumvent high-dimensional integrals for marginal likelihood approximation. We establish consistency and optimality of our method with endogenous application usage. We apply our method to develop personalized prompt schedules in 294 application users, with a goal to maximize the prompt response rate given past application usage and other contextual factors. We found the best push schedule given the same covariates varied among the users, thus calling for personalized policies. Using the estimated personalized policies would have achieved a mean prompt response rate of 23% in these users at 16 weeks or later: this is a remarkable improvement on the observed rate (11%), while the literature suggests 3%-15% user engagement at 3 months after download. The proposed method compares favorably to existing estimation methods including using the R function glmer in a simulation study. In the second part of this dissertation, we aim to solve a practical problem in the mHealth area. Low response rate has been a major issue that blocks researchers from collecting high quality mHealth data. Therefore, developing a prompting system is important to keep user engagement and increase response rate. We aim to learn personalized prompting time for users in order to gain a high response rate. An extension of the personalized learning algorithm is applied on the Intellicare data that incorporates penalties of the population effect parameters and personalized effect parameters into learning the personalized decision rule of sending prompts. The number of personalized policy parameters increases with sample size. Since there is a large number of users in the Intellicare data, it is challenging to estimate such high dimensional parameters. To solve the computational issue, we employ a bagging method that first bootstraps subsamples and then ensembles parameters learned from each subsample. The analysis of Intellicare data shows that sending prompts at a personalized hour helps achieve a higher response rate compared to a one-fit-to-all prompting hour.

El presente trabajo de investigación consiste en crear una idea de negocio innovadora que cubra una necesidad actual y pueda ser viable de ejecutarse. En la investigación se identificó a través de entrevistas de profundidad que existen padres que cuando van a comprar ropa para sus bebés entre cero y dos años, se sienten abrumados porque los bebés se aburren rápidamente. Por esta razón, nace WAWA STORE, una idea de negocio que consiste en fabricar y comercializar ropa personalizada para bebés a través de una Web App. Para el desarrollo del presente proyecto se cuenta con un equipo de trabajo con experiencia en el rubro, uno de los integrantes en la actualidad cuenta con un negocio de ropa para niños y el equipo que lo acomp0aña cuenta con experiencia en administración, finanzas, y ventas. El presente proyecto cuenta con el potencial de generar una rentabilidad promedio anual de 15% asimismo, les permitirá a los accionistas recuperar su inversión con un excedente de S/.139,290 que logra una tasa de rendimiento interna de retorno de 84%. Por último, para poner en marcha el proyecto se requiere una inversión total de S/.89,186 soles, de los cuales se cuenta con el 66% de capital propio. En consecuencia, se busca adquirir un aporte de inversión de S/.30,000 para dar inicio a las operaciones. / The present research work is to create an innovative business idea that meets a current need and can be viable to execute. The research identified through in-depth interviews that there are parents who when they go to buy clothes for their babies between zero and two years old, they feel overwhelmed because babies get bored quickly. For this reason, WAWA STORE was born as a business idea that consists of manufacturing and marketing personalized baby clothes through a web app. For the development of this project there is a work team with experience in the field, one of the members currently has a business of clothing for children and the team that accompanies him has experience in administration, finance and sales. This project has the potential to generate an average annual return of 15% also, it will allow shareholders to recover their investment with a surplus of S/. 139,290 achieving an internal return rate of 89%. Finally, the implementation of the project requires a total investment of S/.89,186 soles, of which 66% own capital is available. Consequently, it seeks to acquire an investment contribution of S/.30,000 to start operations. / Trabajo de investigación

Response-Adaptive (RA) designs are used to adaptively allocate patients in clinical trials. These methods have been generalized to include Covariate-Adjusted Response-Adaptive (CARA) designs, which adjust treatment assignments for a set of covariates while maintaining features of the RA designs. Challenges may arise in multi-center trials if differential treatment responses and/or effects among sites exist. We propose Site-Adjusted Response-Adaptive (SARA) approaches to account for inter-center variability in treatment response and/or effectiveness, including either a fixed site effect or both random site and treatment-by-site interaction effects to calculate conditional probabilities. These success probabilities are used to update assignment probabilities for allocating patients between treatment groups as subjects accrue. Both frequentist and Bayesian models are considered. Treatment differences could also be attributed to differences in social determinants of health (SDH) that often manifest, especially if unmeasured, as spatial heterogeneity amongst the patient population. In these cases, patient residential location can be used as a proxy for these difficult to measure SDH. We propose the Location-Adjusted Response-Adaptive (LARA) approach to account for location-based variability in both treatment response and/or effectiveness. A Bayesian low-rank kriging model will interpolate spatially-varying joint treatment random effects to calculate the conditional probabilities of success, utilizing patient outcomes, treatment assignments and residential information. We compare the proposed methods with several existing allocation strategies that ignore site for a variety of scenarios where treatment success probabilities vary.

Altered signaling pathways, which are mediated by post-translational modifications and changes in protein expression levels, are key regulators of cancer initiation, progression, and therapeutic escape. Many aspects of cancer progression, including early carcinogenesis and immediate response to drug treatment, are beyond the scope of genetic profiling and non-invasive monitoring techniques. Global protein profiling of cancer cell line models, tumor tissues, and biofluids (e.g. serum or urine) using mass spectrometry-based proteomics produces novel biological insights, which support improved patient outcomes. Recent technological advances resulting in next-generation mass spectrometry instrumentation and improved bioinformatics workflows have led to unprecedented measurement reproducibility as well as increased depth and coverage of the human proteome. It is now possible to interrogate the cancer proteome with quantitative proteomics to identify prognostic cancer biomarkers, stratify patients for treatment, identify new therapeutic targets, and elucidate drug resistance mechanisms. There are, however, numerous challenges associated with protein measurements. Biological samples have a high level of complexity and wide dynamic range, which is even more pronounced in samples used for non-invasive disease monitoring, such as serum. Cancer biomarkers are generally found in low abundance compared to other serum proteins, particularly at early stages of disease where cancer detection would make the biggest impact on improving patient survival. Additionally, the large-scale datasets generally require bioinformatics expertise to produce useful biological insights. These difficulties converge to create obstacles for down-stream clinical translation. This dissertation research demonstrates how proteomics is applied to develop new resources and generate novel workflows to improve protein quantification in complex biosamples, which could improve translation of cancer research to benefit patient care. The studies described in this dissertation move from assessment of quantitative mass spectrometry platforms, to analytical assay development and validation, and ending with personalized biomarker development applied to patient samples. As an example, four different quantitative mass spectrometry acquisition platforms are explored and comparisons of their ability to quantify low abundance peptides in a complex background are explored. Lung cancers frequently have aberrant signaling resulting in increased kinase activity and targetable signaling hubs; kinase inhibitors have been successfully developed and implemented clinically. Therefore, changes in amounts of kinase peptides in the complex background of peptides from all ATP-utilizing enzymes in a lung cancer cell line model after kinase inhibitor treatment was selected as a model system. Traditional mass spectrometry platforms, data dependent acquisition and multiple reaction monitoring, are compared to the two newer methods, data independent acquisition and parallel reaction monitoring. Relative quantification is performed across the four methods and analytical performance as well as downstream applications, including drug target identification and elucidation of signaling changes. Liquid chromatography – multiple reaction monitoring (LC-MRM) was selected for development of multiplexed quantitative assays based on superior sensitivity and fast analysis times, allowing for larger peptide panels. Method comparison results also provide guidelines for quantitative proteomics platform selection for translational cancer researchers. Next, a multiplexed quantitative LC-MRM assay targeting a panel of 30 RAS signaling proteins was developed and described. Over 30% of all human cancers have a RAS mutation and these cancers are generally aggressive and limited treatment options, leading to poor patient prognosis. Many targeted inhibitors have successfully shut down RAS signaling, leading to tumor regression, however, acquired drug resistance is common. The multiplexed LC-MRM assays characterized and validated are a publically available resource for cancer researchers to interrogate the RAS signal transduction network. Feasibility has been demonstrated in cell line models in order to identify signaling changes that confer BRAF inhibitor resistance and biomarkers of sensitivity to treatment. This analytical LC-MRM panel could support meaningful development of new therapeutic options and identification of companion biomarkers, with the end goal of improving patient outcomes. Multiplexed LC-MRM assays developed for personalized disease biomarkers using an integrated multi-omics approach are described for Multiple Myeloma, an incurable malignancy with poor patient outcomes. This disease is characterized by clonal expansion of the plasma cells in the bone marrow, which secrete a monoclonal immunoglobulin, or M-protein. Clinical treatment decisions are based on multiple semi-quantitative assays that require manual evaluation. In the clinic, minimal residual disease quantification methods, including multi-parameter flow cytometry and immunohistochemistry, are applied to bone marrow aspirates, which is a highly invasive technique that does not provide a systemic evaluation of the disease. To address these issues, we hypothesized that unique variable region peptides could be identified and LC-MRM assays developed specific to each patient’s M-protein to improve specificity and sensitivity in non-invasive disease monitoring. A proteogenomics approach was used to design personalized assays for each patient to monitor their disease progression, which demonstrate improved specificity and up to a 500-fold increase in sensitivity compared to current clinical methods. Assays can be developed from marrow aspirates collected when the patient was at residual disease stage, which is useful if no sample with high disease burden is available. The patient-specific tests are also multiplexed with constant region peptide assays that monitor all immunoglobulin heavy and light chain classes, which could reduce analysis to a single test. In conclusion, highly sensitive patient-specific assays have been developed that could change the paradigm for patient evaluation and clinical decision-making, increasing the ability of clinicians to continue first line therapy in the hopes of achieving a cure, or to intervene at an earlier time point in disease recurrence. This study also provides a blueprint for future development of personalized diagnostics, which could be applied to biomarkers of other cancer types. Overall, these studies demonstrate how quantitative proteomics can be used to support translational cancer research, from the impact of different mass spectrometry platforms on elucidating signaling changes and drug targets to the characterization of multiplexed LC-MRM assays applied to cell line models for translational research purposes and in patient serum samples optimized for clinical translation. We believe that mass spectrometry-based proteomics is poised to play a pivotal role in personalized diagnostics to support implementation of precision medicine, an effort that will improve the quality and efficiency of patient care.

Contextual information retrieval (CIR) is a critical technique for today’s search engines in terms of facilitating queries and returning relevant information. Despite its importance, little progress has been made in its application, due to the difficulty of capturing and representing contextual information about users. This thesis details the development and evaluation of the contextual SERL search, designed to tackle some of the challenges associated with CIR from the World Wide Web. The contextual SERL search utilises a rich contextual model that exploits implicit and explicit data to modify queries to more accurately reflect the user’s interests as well as to continually build the user’s contextual profile and a shared contextual knowledge base. These profiles are used to filter results from a standard search engine to improve the relevance of the pages displayed to the user. The contextual SERL search has been tested in an observational study that has captured both qualitative and quantitative data about the ability of the framework to improve the user’s web search experience. A total of 30 subjects, with different levels of search experience, participated in the observational study experiment. The results demonstrate that when the contextual profile and the shared contextual knowledge base are used, the contextual SERL search improves search effectiveness, efficiency and subjective satisfaction. The effectiveness improves as subjects have actually entered fewer queries to reach the target information in comparison to the contemporary search engine. In the case of a particularly complex search task, the efficiency improves as subjects have browsed fewer hits, visited fewer URLs, made fewer clicks and have taken less time to reach the target information when compared to the contemporary search engine. Finally, subjects have expressed a higher degree of satisfaction on the quality of contextual support when using the shared contextual knowledge base in comparison to using their contextual profile. These results suggest that integration of a user’s contextual factors and information seeking behaviours are very important for successful development of the CIR framework. It is believed that this framework and other similar projects will help provide the basis for the next generation of contextual information retrieval from the Web.