Wolfgang Stille

A user-aware tour proposal framework using a hybrid optimization approach

This paper considers a context- and user-aware approach to the tour planning in Geographic Information Systems. We regard the following scenario: a person with specific preferences and a fixed amount of time is visiting a foreign place of interest. Therefore spatial as well as user specific information is modeled as a multimodally attributed digraph structure, where nodes represent the points of interest and arcs the user optimal paths between them. The goal is to find a set of nodes and arcs representing a round trip that is best satisfying the priorities of the user. This results in an Enhanced Profitable Tour Problem (EPTP), which is - as a generalization of the well-known Traveling Salesman Problem (TSP) - an NP-hard optimization problem. In this article, we introduce an efficient algorithm for this problem and discuss the modeling and implementation details.

Semantically faceted navigation with topic pies

Faceted search allows navigating through large collections along different dimensions in order to find relevant objects efficiently. Traditional faceted search systems often suffer from a lack of usability; furthermore facets are often static and independent from the search result set. In this paper, we present a dynamic semantic topical faceting approach. It uses a pie menu called topic pie that allows visualisation of facets and user interaction. Depending on the search query, the topic pie presents a set of topics and major topics which help the user to drill down the search result set to relevant objects efficiently as well as to browse exploratively through the collection. The underlying algorithm optimises the conflicting goals relevance and diversity while avoiding information overload. It reveals a good performance on large data sets. As our use-case, we chose literature research in scientific libraries. An evaluation shows major advantages of our approach compared to state-of-the-art faceted search techniques in nowadays library portals.

Propagating Maximum Capacities for Recommendation

Neighborhood-based approaches often fail in sparse scenarios; a direct implication for recommender systems exploiting co-occurring items is often an inappropriately poor performance. As a remedy, we propose to propagate information (e.g., similarities) across the item graph to leverage sparse data. Instead of processing only directly connected items (e.g. co-occurrences), the similarity of two items is defined as the maximum capacity path interconnecting them. Our approach resembles a generalization of neighborhood-based methods that are obtained as special cases when restricting path lengths to one. We present two efficient online computation schemes and report on empirical results.