Tim Hussein

Hybreed: A software framework for developing context-aware hybrid recommender systems

This article introduces Hybreed, a software framework for building complex context-aware applications, together with a set of components that are specifically targeted at developing hybrid, context-aware recommender systems. Hybreed is based on a concept for processing context that we call dynamic contextualization. The underlying notion of context is very generic, enabling application developers to exploit sensor-based physical factors as well as factors derived from user models or user interaction. This approach is well aligned with context definitions that emphasize the dynamic and activity-oriented nature of context. As an extension of the generic framework, we describe Hybreed RecViews, a set of components facilitating the development of context-aware and hybrid recommender systems. With Hybreed and RecViews, developers can rapidly develop context-aware applications that generate recommendations for both individual users and groups. The framework provides a range of recommendation algorithms and strategies for producing group recommendations as well as templates for combining different methods into hybrid recommenders. Hybreed also provides means for integrating existing user or product data from external sources such as social networks. It combines aspects known from context processing frameworks with features of state-of-the-art recommender system frameworks, aspects that have been addressed only separately in previous research. To our knowledge, Hybreed is the first framework to cover all these aspects in an integrated manner. To evaluate the framework and its conceptual foundation, we verified its capabilities in three different use cases. The evaluation also comprises a comparative assessment of Hybreed’s functional features, a comparison to existing frameworks, and a user study assessing its usability for developers. The results of this study indicate that Hybreed is intuitive to use and extend by developers.

Choice-based preference elicitation for collaborative filtering recommender systems

We present an approach to interactive recommending that combines the advantages of algorithmic techniques with the benefits of user-controlled, interactive exploration in a novel manner. The method extracts latent factors from a matrix of user rating data as commonly used in Collaborative Filtering, and generates dialogs in which the user iteratively chooses between two sets of sample items. Samples are chosen by the system for low and high values of each latent factor considered. The method positions the user in the latent factor space with few interaction steps, and finally selects items near the user position as recommendations. In a user study, we compare the system with three alternative approaches including manual search and automatic recommending. The results show significant advantages of our approach over the three competing alternatives in 15 out of 24 possible parameter comparisons, in particular with respect to item fit, interaction effort and user control. The findings corroborate our assumption that the proposed method achieves a good trade-off between automated and interactive functions in recommender systems.

MODELING AND EXPLOITING CONTEXT FOR ADAPTIVE COLLABORATION

Collaborative work is characterized by frequently changing situations and corresponding demands for tool support and interaction behavior provided by the collaboration environment. Current approaches to address these changing demands include manual tailoring by end-users and automatic adaptation of single user tools or for individual users. Few systems use context as a basis for adapting collaborative work environments, mostly focusing on document recommendation and awareness provision. In this paper, we present, firstly, a generic four layer framework for modeling and exploiting context. Secondly, a generic adaptation process translating user activity into state, deriving context for a given focus, and executing adaptation rules on this context. Thirdly, a collaboration domain model for describing collaboration environments and collaborative situations. Fourthly, examples of exploiting our approach to support context-based adaptation in four typical collaboration situations: co-location, co-access, co-recommendation, and co-dependency.

Semantic Models for Adaptive Interactive Systems

Providing insights into methodologies for designing adaptive systems based on semantic data, and introducing semantic models that can be used for building interactive systems, this book showcases many of the applications made possible by the use of semantic models. Ontologies may enhance the functional coverage of an interactive system as well as its visualization and interaction capabilities in various ways. Semantic models can also contribute to bridging gaps; for example, between user models, context-aware interfaces, and model-driven UI generation. There is considerable potential for using semantic models as a basis for adaptive interactive systems. A variety of reasoning and machine learning techniques exist that can be employed to achieve adaptive system behavior. The advent and rapid growth of Linked Open Data as a large-scale collection of semantic data has also paved the way for a new breed of intelligent, knowledge-intensive applications. Semantic Models for Adaptive Interactive Systems includes ten complementary chapters written by experts from both industry and academia. Rounded off by a number of case studies in real world application domains, this book will serve as a valuable reference for researchers and practitioners exploring the use of semantic models within HCI.

A framework and an architecture for context-aware group recommendations

In this paper, we propose a generic framework to generate context-aware recommendations for both single users as well as groups. We present the the concept of context views and a corresponding architecture implementing the framework as well as exemplary recommendation workflows for group recommendations.

Explanation of spreading activation based recommendations

In this paper, we introduce an approach for explaining recommendations in environments that are based on semantic models. Using a constrained Spreading Activation (CSA) technique for recommendation generation, we store and exploit the activation paths leading to recommendations. These paths later can be used to generate both verbal explanations and relevance feedback forms.

Generating route instructions with varying levels of detail

In this paper, we present a technique for adaptive generation of personalized route instructions based on the drivers knowledge of particular route sections. We evaluated the mechanism with two empirical studies, both attesting significant preference for the adaptively generated presentations over an established online service (Google Maps).

Semi-automatic generation of recommendation processes and their GUIs

Creating and optimizing content- and dialogue-based recommendation processes and their GUIs (graphical user interfaces) manually is expensive and slow. Changes in the environment may also be found too late or even be overlooked by humans. We show how to generate such processes and their GUIs semi-automatically by using knowledge derived from unstructured data such as customer feedback on products on the Web. Our approach covers the whole lifecycle from knowledge discovery through text mining techniques to the use of this knowledge for semi-automatic generation of recommendation processes and their user interfaces as well as their comparison in real-world use within the e-commerce domain through A/B-variant tests. These tests indicate that our approach can lead to better results as well as less manual effort.

A Context-Aware Shopping Portal Based on Semantic Models

This chapter illustrates how semantic models can be used as a backend data source for both exploration and adaptation purposes. For a fictitious shopping portal, we implemented a faceted navigation approach that provides means for exploring the portal’s content manually. In addition to that, we implemented an adaptation mechanism based on spreading activation that also exploits the semantic structure of the underlying data.

Service-based recommendations for context-aware navigation support

The whole automotive industry is changing rapidly as fossil fuel is getting more expensive, pollution is a serious problem, and people often just expect more from their vehicle than “only” getting them from A to B. Electric mobility is a promising answer for the first two questions. Unfortunately, it comes with some serious drawbacks: Charging a battery takes up to several hours (compared to a few minutes for refueling) while cruising ranges are short. As a result, intelligent routing systems will be even more important in the future, in order to avoid long and boring waiting times at charging points (which might perhaps also be unnecessarily expensive). If the driver has to wait for such a long time, he should be guided to a charging station where he can find attractive services and shopping or entertainment opportunities matching his interests – ideally, at a charging point with reasonable prices. Apart from electric mobility, such navigation solutions can be attractive in many other scenarios, if providing added-value services. For instance could they recommend certain hotels along the route, music matching the driver’s taste or simply a garage with certain bargains such as reduced winter tyres, if the old ones are used up.