No description available.

No description available.

No description available.

Användningen och behovet av rekommendationssystem i digitala tjänster har växt i takt med att utbudet i dessa blivit allt större och svårare för användare att navigera i. Rekommendationssystem används idag i allt ifrån E-handel till musikoch filmstreaming. För att förse användare med rekommendationer på objekt används en mängd olika väl beprövade algoritmer, filtreringsmetoder och datainsamlingsmetoder. Att applicera dessa i livestreamingtjänster ställer nya krav på systemen eftersom innehållet byts ut mer frekvent, helt nytt innehåll tillkommer regelbundet och explicit data samt metadata är sällan tillräcklig för att ta fram träffsäkra rekommendationer. I en fallstudie med företaget Liveguide undersöks hur rekommendationssystem kan appliceras i livestreamingtjänster med avseende på de utmaningar och krav som finns. Metoder presenteras där aktuella lösningar testas, utvärderas och anpassas till att fungera bra i livestreamingsammanhang. Slutligen föreslås tre modeller för rekommendationssystem som tagits fram utifrån det resultat metoderna leder till. För att tillfredsställa de identifierade utmaningarna inom området visade sig hybrida, mångsidiga modeller fördelaktiga i livestreaming. / The usage and demand of recommender systems in digital services has increased in line with their huge range of products, making it more difficult for users to navigate through the content. Recommender systems are used in a wide scope of digital services ranging from E-commerce to music and film streaming. In order to provide users with recommendations on objects, a variety of algorithms, filtering methods and methods of data collections are being used. Applying these in live streaming services puts new demands on such systems since the content is replaced frequently and new objects added regularly. Furthermore, livestreaming services often lack explicit data and metadata, making recommendations less accurate. In a case study with Liveguide, recommender systems are evaluated, focusing on whether they are applicable to live streaming services, respecting requirements and demands on such systems. Methods are presented which tests, evaluates and adapts existing solutions to fit in well in context of live streaming. Finally, three models for recommender systems are suggested, based on the methods result. In order to satisfy the identified challenges, hybrid models turned out to be preferable in the context.

Grid environments are inherently heterogeneous. If the computational power provided by collaborations on the Grid is to be harnessed in the true sense, there is a need for applications that can automatically adapt to changes in the execution environment. The application writer should not be burdened with the job of choosing the right algorithm and implementation every time the resources on which the application runs are changed. A lot of research has been done in adapting applications to changing conditions. The existing systems do not address the issue of providing a unified interface to permit algorithm selection at runtime. The goal of this research is to design and develop a unified interface to applications in order to permit seamless access to different algorithms providing similar functionalities. Long running, computationally intensive scientific applications can produce huge amounts of performance data. Often, this data is discarded once the application's execution is complete. This data can be utilized in extracting information about algorithms and their performance. This information can be used to choose algorithms intelligently. The research described in this thesis aims at designing and developing a component based unified interface for runtime algorithm selection in grid environments. This unified interface is necessary so that the application code does not change if a new algorithm is used to solve the problem. The overhead associated with making the algorithm choice transparent to the application is evaluated. We use a data mining approach to algorithm selection and evaluate its potential effectiveness for scientific applications. / Master of Science

Institutional support of higher teaching and learning at scale poses three unique challenges. The first challenge is poor institutional accounting of instructors' use of educational platforms and software, especially the learning management system (LMS). The second challenge is a deficit of trust among stakeholders with unique job roles, prerogatives, and editorial preferences. The third challenge is one-size-fits-all, open-loop, or stopgap support processes. To address these challenges, this three-phase dissertation project proposes a novel sociotechnical framework for institutional support using trustworthy educational recommender systems. This framework accounts for LMS platform contexts, multiple stakeholders, and editorial trust relationships. In its first phase, the project proposes ``Depth of Use" (DOU): a first-principles framework of frequent LMS use-contexts. DOU is found to highlight low-adoption course cohorts, evaluate course design interventions, and improve IT emergency preparedness. The second phase of this project proposes a novel model of recommendation trustworthiness based in stakeholder allocation of RS editorial tasks. The study discovers a spectrum of faculty intentions about editorial division-of-labor and its frequent rationales, including student expertise, professional curriculum needs, authorship burdens at scale, and learner disengagement. In its third phase, the project investigates how editorial trust might be enhanced by transparency cues (guarantees, social proof, content tags). The dissertation concludes with a set of design guidelines to aid HCI practitioners in enhancing editorial transparency and algorithmic explainability, and increasing process efficacy of institutional support. / Doctor of Philosophy / In higher education, supporting faculty effectively can be challenging, especially with technology use at scale. This dissertation reckons with three primary aspects of this challenge: inadequate tracking of how educators use learning platforms, low trust among different institutional stakeholders, and inefficient support processes. To address these challenges, we propose a novel framework to personalize instructional support using learning management systems (LMS) as platforms to reach out to faculty, interpret their technology needs, and deliver interventions using educational recommender systems (ERS). Our framework allows better understanding of faculty's LMS use, editorial intent, and trust of automation. It also highlights structural barriers to trust and process efficacy at universities. Finally, it delivers guidelines for the design of trustworthy educational recommendation and support processes.

In recent years the market share of major real estate companies, at least the Czech ones, has not changed much. Statistical data don't reflect any significal upward trend in volumes of properties for rent or sale. In case the real estate company would like to access larger market share, they have to secure a competitive advantage over the others. One of the ways how to attract more potential customers might be speeding up the company website's property search process. In many cases the website visitors are facing hundreds or thousands of property offers before finding couple satisfactories. The aim of the thesis is to explore possibilities of applicating customer preferences in property trading. The focus is put on research of recommender system algorithms, their characteristics and limtations. The author is evaluating usage of each algorithm variant and its suitability for a real world deployment in a real estate area. Apart from the theoretical part of the work one can find a part, where real estate information system is extended with a framework for implementing recommendation system algorithms. The author is in possesion of production data of a medium sized real estate company. He uses the recommender system framework to build and evaluate example algorithm. Powered by TCPDF (www.tcpdf.org)

[pt] Sistemas de recomendação têm sido amplamente utilizados pelos grandes portais na Web, em decorrência do aumento do volume de dados disponíveis na Web. Tais sistemas são basicamente utilizados para sugerir informações relevantes para os seus usuários. Esta dissertação apresenta um sistema de recomendação de segundo nível para auxiliar equipes de jornalistas de portais de notícias no processo de recomendação de notícias relacionadas para os usuários do portal. O sistema é chamado de segundo nível pois apresenta recomendações aos jornalistas para que, por sua vez, geram recomendações aos usuários do portal. O modelo seguido pelo sistema consiste na recomendação de notícias relacionadas com base em características extraídas do próprio texto da notícia original. As características extraídas permitem a criação de consultas contra um banco de dados de notícias anteriormente publicadas. O resultado de uma consulta é uma lista de notícias candidatas à recomendação, ordenada pela similaridade com a notícia original e pela data de publicação, que o editor da notícia original manualmente processa para gerar a lista final de notícias relacionadas. / [en] Recommendation systems are widely used by major Web portals due to the increase in the volume of data available on the Web. Such systems are basically used to suggest information relevant to their users. This dissertation presents a second-level recommendation system, which aims at assisting the team of journalists of a news Web portal in the process of recommending related news for the users of the Web portal. The system is called second level since it creates recommendations to the journalists Who, in turn, generate recommendations to the users. The system follows a model based on features extracted from the text itself. The extracted features permit creating queries against a news database. The query result is a list of candidate news, sorted by score and date of publication, which the news editor manually processes to generate the final list of related news.

Avec la quantité croissante du contenu pédagogique produit chaque jour par les utilisateurs, il devient très difficile pour les apprenants de trouver les ressources les plus adaptées à leurs besoins. Les systèmes de recommandation sont utilisés dans les plateformes éducatives pour résoudre le problème de surcharge d'information. Ils sont conçus pour fournir des ressources pertinentes à un apprenant en utilisant certaines informations sur les utilisateurs et les ressources. Le présent travail s'inscrit dans le contexte des systèmes de recommandation des ressources pédagogiques, en particulier les systèmes qui utilisent des informations sociales. Nous avons défini une approche de recommandation de ressources éducatives en se basant sur les résultats de recherche dans le domaine des systèmes de recommandation, des réseaux sociaux et des environnements informatiques pour l’apprentissage humain. Nous nous appuyons sur les relations sociales entre apprenants pour améliorer la précision des recommandations. Notre proposition est basée sur des modèles formels qui calculent la similarité entre les utilisateurs d'un environnement d'apprentissage pour générer trois types de recommandation, à savoir la recommandation des 1) ressources populaires, 2) ressources utiles et 3) ressources récemment consultées. Nous avons développé une plateforme d'apprentissage, appelée Icraa, qui intègre nos modèles de recommandation. La plateforme Icraa est un environnement d’apprentissage social qui permet aux apprenants de télécharger, de visualiser et d’évaluer les ressources éducatives. Dans cette thèse, nous présentons les résultats d'une expérimentation menée pendant deux ans qui a impliqué un groupe de 372 apprenants d'Icraa dans un contexte éducatif réel. L'objectif de cette expérimentation est de mesurer la pertinence, la qualité et l'utilité des ressources recommandées. Cette étude nous a permis d'analyser les retours des utilisateurs concernant les trois types de recommandations. Cette analyse a été basée sur les traces des utilisateurs enregistrées avec Icraa et sur un questionnaire. Nous avons également effectué une analyse hors ligne en utilisant un jeu de données afin de comparer notre approche avec quatre algorithmes de référence. / With the increasing amount of educational content produced daily by users, it becomes very difficult for learners to find the resources that are best suited to their needs. Recommendation systems are used in educational platforms to solve the problem of information overload. They are designed to provide relevant resources to a learner using some information about users and resources. The present work fits in the context of recommender systems for educational resources, especially systems that use social information. We have defined an educational resource recommendation approach based on research findings in the area of recommender systems, social networks, and Technology-Enhanced Learning. We rely on social relations between learners to improve the accuracy of recommendations. Our proposal is based on formal models that calculate the similarity between users of a learning environment to generate three types of recommendation, namely the recommendation of 1) popular resources; 2) useful resources; and 3) resources recently consulted. We have developed a learning platform, called Icraa, which integrates our recommendation models. The Icraa platform is a social learning environment that allows learners to download, view and evaluate educational resources. In this thesis, we present the results of an experiment conducted for almost two years on a group of 372 learners of Icraa in a real educational context. The objective of this experiment is to measure the relevance, quality and usefulness of the recommended resources. This study allowed us to analyze the user’s feedback on the three types of recommendations. This analysis is based on the users’ traces which was saved with Icraa and on a questionnaire. We have also performed an offline analysis using a dataset to compare our approach with four base line algorithms.

Sistemas de Recomendação (SR) vêm se apresentando como poderosas ferramentas para portais web tais como sítios de comércio eletrônico. Para fazer suas recomendações, os SR se utilizam de fontes de dados variadas, as quais capturam as características dos usuários, dos itens e suas transações, bem como de modelos de predição. Dada a grande quantidade de dados envolvidos, é improvável que todas as recomendações possam ser bem representadas por um único modelo global de predição. Um outro importante aspecto a ser observado é o problema conhecido por cold-start, que apesar dos avanços na área de SR, é ainda uma questão relevante que merece uma maior atenção. O problema está relacionado com a falta de informação prévia sobre novos usuários ou novos itens do sistema. Esta tese apresenta uma abordagem híbrida de recomendação capaz de lidar com situações extremas de cold-start. A abordagem foi desenvolvida com base no algoritmo SCOAL (Simultaneous Co-Clustering and Learning). Na sua versão original, baseada em múltiplos modelos lineares de predição, o algoritmo SCOAL mostrou-se eficiente e versátil, podendo ser utilizado numa ampla gama de problemas de classificação e/ou regressão. Para melhorar o algoritmo SCOAL no sentido de deixá-lo mais versátil por meio do uso de modelos não lineares, esta tese apresenta uma variante do algoritmo SCOAL que utiliza modelos de predição baseados em Máquinas de Aprendizado Extremo. Além da capacidade de predição, um outro fator que deve ser levado em consideração no desenvolvimento de SR é a escalabilidade do sistema. Neste sentido, foi desenvolvida uma versão paralela do algoritmo SCOAL baseada em OpenMP, que minimiza o tempo envolvido no cálculo dos modelos de predição. Experimentos computacionais controlados, por meio de bases de dados amplamente usadas na prática, comprovam que todos os desenvolvimentos propostos tornam o SCOAL ainda mais atraente para aplicações práticas variadas. / Recommender Systems (RS) are powerful and popular tools for e-commerce. To build its recommendations, RS make use of multiple data sources, capture the characteristics of items, users and their transactions, and take advantage of prediction models. Given the large amount of data involved in the predictions made by RS, is unlikely that all predictions can be well represented by a single global model. Another important aspect to note is the problem known as cold-start that, despite that recent advances in the RS area, it is still a relevant issue that deserves further attention. The problem arises due to the lack of prior information about new users and new items. This thesis presents a hybrid recommendation approach that addresses the (pure) cold start problem, where no collaborative information (ratings) is available for new users. The approach is based on an existing algorithm, named SCOAL (Simultaneous Co-Clustering and Learning). In its original version, based on multiple linear prediction models, the SCOAL algorithm has shown to be efficient and versatile. In addition, it can be used in a wide range of problems of classification and / or regression. The SCOAL algorithm showed impressive results with the use of linear prediction models, but there is still room for improvements with nonlinear models. From this perspective, this thesis presents a variant of the SCOAL based on Extreme Learning Machines. Besides improving the accuracy, another important issue related to the development of RS is system scalability. In this sense, a parallel version of the SCOAL, based on OpenMP, was developed, aimed at minimizing the computational cost involved as prediction models are learned. Experiments using real-world datasets has shown that all proposed developments make SCOAL algorithm even more attractive for a variety of practical applications.