Christian Voigtmann

An approach to Collaborative Context Prediction

Context prediction approaches forecast future contexts based on known context patterns to adapt e.g., services in advance. In the case of the users context history not providing suitable context information for the observed context pattern, to the best of our knowledge context prediction algorithms will fail to forecast the appropriate future context. To overcome the gap of missing context information in the users context history, we propose the Collaborative Context Prediction (CCP) approach. CCP utilises the collaborative characteristics of existing recommendation systems of social networks. To evaluate the CCP method an experimental comparison of the proposed method against the local Alignment context predictor is carried out.

Enhancing alignment based context prediction by using multiple context sources: experiment and analysis

Context aware applications are reactive, they adapt to an entitys context when the context has changed. In order to become proactive and act before the context actually has changed future contexts have to be predicted. This will enable functionalities like preloading of content or detection of future conflicts. For example if an application can predict where a user is heading to it can also check for train delays on the users way. So far research concentrates on context prediction algorithms that only use a history of one context to predict the future context. In this paper we propose a novel multidimensional context prediction algorithm and we show that the use of multidimensional context histories increases the prediction accuracy. We compare two multidimensional prediction algorithms, one of which is a new approach; the other was not yet experimentally tested. In theory, simulation and a real world experiment we verify the feasibility of both algorithms and show that our new approach has at least equal or better reasoning accuracy.

A new approach for distributed and collaborative context prediction

The processing capabilities of current smartphones have increased significantly. We propose a distributed and collaborative context prediction approach that exclusively uses current smartphones to automatically collect, process and predict contexts of users. To predict a users next context, not only her context history is utilised but also context histories of other users are used. The communication between the smartphones of the users is realised using peer-2-peer. Therefore, no centralised server unit is needed to process the context information of the users externally. We provide a proof-of-concept implementation and present experimental results that demonstrate the practicality of the proposed architecture.

Prospects for Context Prediction Despite the Principle of Informational Self-Determination

Until now informational self-determination has not been examined in relation to context prediction. The right of users to decide whether or not and how they want their personal data to be used is not ensured. This paper analyses the problem by introducing the state of the art and application fields of context prediction with focus on the users’ personal data. Subsequently, informational self-determination and the problems that arise if techniques of context prediction compete with informational self determination will be presented. Finally, possible solutions for overcoming these problems will be outlined.

Datenschutzgerechte Gestaltung von kontextvorhersagenden Algorithmen

ZusammenfassungDie Vorhersage von Bedürfnissen, Interessen und ähnlichen Informationen ist eine Grundfunktion künftiger IT-Anwendungen. Im Rahmen des interdisziplinären VENUS-Projekts wurde untersucht, wie sich entsprechende Funktionen datenschutzrechtlich auswirken. Daraus leiten die Autoren konkrete Vorschläge zu ihrer datenschutzgerechten Gestaltung ab.

Support-U: Designing an Ambient Assisted Living System Using Interdisciplinary Development Patterns

Software development has proven to be a challenge. To address this challenge, there are quite many interesting approaches how to develop software—starting from the waterfall approach, up to recently quite popular agile software development techniques. Another already some years old approach was described by the Gang of Four and proposes the usage of design patterns to provide a general reusable solution to commonly occurring problems in software development. Although design patterns have been around for a long time, their usability is still promising. To the best of our knowledge “interdisciplinary patterns” to address challenges in the development of context aware application in ubiquitous environments have not been described in literature so far. Hence, this chapter proposes and also evaluates concrete interdisciplinary software development patterns. To provide an application example the proposed patterns are used to address two use cases that commonly occur in the development process of context aware applications: providing transparency to the user and ensuring a user’s self-determination. For the demonstration of the patterns Support-U a context aware application that provides elderly people to live autonomously is used.

Collaborative Context Prediction

Context prediction is used to proactively adapt e.g., services to users’ needs. Due to the fact that context prediction enables proactiveness it has a high significance for UC systems. To the best of our knowledge, research literature on context prediction only focuses on the history of the user whose next context has to be predicted. Does a user suddenly change her behaviour in an unexpected way, the context history of the user does not contain appropriate context information to provide reliable context predictions. Hence, context prediction algorithms will fail to predict a user’s next context if they solely rely on the context history of the user, whose context has to be predicted. To overcome the gap of missing context information in the user’s context history, the Collaborative Context Prediction (CCP) approach is proposed. CCP takes advantage of existing direct and indirect relations which may exist among the context histories of various users. Thereby, CCP bases on the Higher-order Singular Value Decomposition, which is also applied in the field of recommendation systems. To provide an evaluation of CCP it is compared to state-of-the-art context prediction approaches with respect to its prediction accuracy using a collaborative data set. For the reason that context prediction approaches primarily use personal context data legal criteria are presented. These criteria are used to legally assess the context prediction approaches. Subsequently, the resulting consequences are discussed.