Logo des Repositoriums

Auflistung nach Schlagwort "recommender systems"

1 - 6 von 6
  • Konferenzbeitrag
    A User Interface Concept for Context-Aware Recommender Systems
    (Mensch und Computer 2016 - Tagungsband, 2016) Hiesel, Patrick; Wörndl, Wolfgang; Braunhofer, Matthias; Herzog, Daniel
    A context-aware recommender system incorporates the knowledge of different contextual factors - such as time or weather information - to improve the item suggestions made to a user. While this provides great benefit to users, it might be hard for them to grasp why certain items are relevant, given the complexity of a context-aware recommender. In this paper, we propose, implement and evaluate a user interface concept that seeks to tackle this challenge. We show how popularity graphs can be used to inform the user about the relevance of items in different contexts and how users perceive different contextual factors given our concept. A user study with 14 participants demonstrates that our concept is valid and appreciated by users.
  • Konferenzbeitrag
    Architecture of a recommender system to support collaboration in a software environment
    (WM 2003: Professionelles Wissesmanagement – Erfahrungen und Visionen, Beiträge der 2. Konferenz Professionelles Wissensmanagement, 2003) Lichtnow, Daniel; Loh, Stanley; Saldana Garin, Ramiro; Caringi, Augusto; Anjos, Pablo Lucas dos
    Within organizations, people learn through exchanging knowledge. This kind of task (named collaboration) is important for the organizational learning. Collaboration can be supported by Information Technology tools as chats, newsgroups, forums and e-mailing lists. However, this kind of support only enables message exchange, lacking to help people in the learning process. This work presents the architecture of a recommender system to support collaboration among people in an software organization. The system analyzes textual messages sent during the session, identifies the context of the discussion and suggests documents, authorities (people with competence in a subject) and past discussions within the same context.
  • Konferenzbeitrag
    Augmented Reality Based Recommending in the Physical World
    (Mensch und Computer 2018 - Workshopband, 2018) Álvarez Márquez, Jesús Omar; Ziegler, Jürgen
    Recommender systems have received the attention of the scientific community for a long time now and they have become a daily tool for internet users. Nonetheless, they are not commonly applied to physical settings, where having access to recommendations could be of great benefit, specially when combined with item comparison capabilities. Due to the latest augmented reality technology advances, it is possible to bring these concepts together. An intuitive action like visually comparing two products could be enhanced by 3D cues and suggestions. In such terms, we discuss the possibilities to improve the item exploration and decision-making stages of the recommending process by providing item comparison supported by 3D augmentations, offering a novel contribution to both augmented reality and recommender systems domains.
  • Workshopbeitrag
    Picture Based Food Recommendation
    (Mensch und Computer 2016 - Tagungsband, 2016) Meißner, Hanns; Stahl, Michael
    Das Teilen von Fotografien aller Art ist ein wichtiger Bestandteil in sozialen Medien geworden. Gerade das Phänomen Foodporn, Fotos von schmackhaft angerichteten Speisen, erlebt in den letzten Jahren auf Plattformen wie Facebook oder Instagram einen regelrechten Boom. Der in diesem Paper vorgestellte Prototyp eines bildbasierten Recommender-Systems nutzt solche Fotos als Input und Output. Der Nutzerinput besteht lediglich aus einem “like” oder “dislike” eines Foodporn-Fotos. Der Recommender soll Rezepte vorschlagen, die dem persönlichen Geschmack und aktuellen Bedürfnis auf Essen des Nutzers entsprechen, indem er nutzerspezifische Vorlieben erlernt, und dabei hilft geeignete Kochideen zu gewinnen. Die Datengrundlage schafft ein Crowdsourcing-Ansatz zum Sammeln von Tags zu den jeweiligen Fotos der in das System integriert ist. Die gängigsten Gamification-Features wurden in einer Community-Komponente eingebaut, um die User für die Mitwirkung am Datensatz zu motivieren. In einer tagebuchähnlichen Langzeitstudie wurde das neuartige Recommender-Konzept untersucht und erfolgreich validiert. Die Gamification-Features erzielten einen positiven Effekt. Künftiges Forschungsinteresse könnten auf der Weiterentwicklung und Evaluation des Recommender-Algorithmus und der Anwendung an sich sowie der Umsetzung als native mobile App liegen.
  • Textdokument
    Reranking-based Recommender System with Deep Learning
    (INFORMATIK 2017, 2017) Saleh, Ahmed; Mai, Florian; Nishioka, Chifumi; Scherp, Ansgar
    An enormous volume of scientific content is published every year. The amount exceeds by far what a scientist can read in her entire life. In order to address this problem, we have developed and empirically evaluated a recommender system for scientific papers based on Twitter postings. In this paper, we improve on the previous work by a reranking approach using Deep Learning. Thus, after a list of top-k recommendations is computed, we rerank the results by employing a neural network to improve the results of the existing recommender system. We present the design of the deep reranking approach and a preliminary evaluation. Our results show that in most cases, the recommendations can be improved using our Deep Learning reranking approach.
  • Konferenzbeitrag
    Web Site Adaptation: Recommendation and Automatic Generation of Navigation Menus
    (12. GI-Workshop "Adapitivität und Benutzermodellierung in interaktiven Softwaresystemen", 2004) Hollink, Vera; van Someren, Maarten; ten Hagen, Stephan
    In recommender systems the interests of users are typically represented as unordered sets of pages. However, on web sites where the pages are not independent of each other, the order in which the pages are visited is important. A recommender should not only recommend the right pages, but also recommend them in the right order. A widely used form of recommending is the context dependent navigation menu. In this short paper we compare content and usage based methods to build navigation menus from scratch. In our experiments the navigation traces of users clearly showed stages, which indicates that users prefer to view pages in a specific order. Content based methods appear to be adequate for creating menus that group of pages with similar topics, but do not provide an ordering. Usage based methods are necessary to recommend pages that match a user’s current navigation stage.