Ralph Kölle

Multilingual Retrieval Experiments with MIMOR at the University of Hildesheim

Fusion and optimization based relevance judgements have proven to be successful strategies in information retrieval. In this year’s CLEF campaign we applied these strategies to multilingual retrieval with four languages. Our fusion experiments were carried out using freely available software. We used the snowball stemmers, internet translation services and the text retrieval tools in Lucene and the new MySQL.

Patent retrieval experiments in the context of the CLEF IP track 2009

At CLEF 2009 the University of Hildesheim submitted experiments for the new Intellectual Property Track. We focused on the main task of this track that aims at finding prior art for a specified patent. Our experiments were split up into one official German run as well as different additional runs using English and German terms. The submitted run was based on a simple baseline approach including stopword elimination, stemming and simple term queries. Furthermore, we investigated the significance of the International Patent Classification (IPC). During the experiments, different parts of a patent were used to construct the queries. In a first stage, only title and claims were included. In contrast, for the post runs we generated a more complex boolean query, which combined terms of the title, claims, description and the IPC classes. The results made clear that using the IPC codes can particularly increase the recall of a patent retrieval system.

Design for emotion: a case study

Increasing the potential of mobile applications to elicit positive emotions can help overcome the prevalent ephemerality of such applications. The current paper addresses the possibilities to support the elicitation of favorable emotions during the development process of mobile applications. It provides a short overview of the theoretical background of emotion research and design to subsequently present an experiential study, focused on the design of positive emotional resonance for an iPhone ride sharing service in terms of desire, satisfaction and fascination.

Domain Specific Retrieval Experiments with MIMOR at the University of Hildesheim

For our first participation in CLEF we chose the domain specific GIRT corpus. We implemented the adaptive fusion model MIMOR (Multiple Indexing and Method-Object Relations) which is based on relevance feedback. The linear combination of several retrieval engines was optimized. As a basic retrieval engine, IRF from NIST was employed. The results are promising. For several topics, our runs achieved a performance above the average. The optimization based on topics and relevance judgements from CLEF 2001 proved to be a fruitful strategy.

Gamification Elements and Their Perceptionby Different Gamer Types

Gamification is being applied more and more to information systems. Many developers apply similar strategies for the gamification of tasks. A benchmark conducted in 22 mobile apps containing gamification elements showed reward points as the most frequent design element. In a pilot user experiment, we analyzed the user perception of three gamification elements including reward points in the context of project management. Reward points are seen as motivating; in contrast a leader board was viewed more critically. It was shown that the judgment of elements depends on the game personality type. The gamer type Killer had a medium and significant correlation with a positive evaluation of the gaming elements. Only the gamer type Achiever had a positive and significant correlation with a positive judgment of rewards points.

Mono-and Crosslingual Retrieval Experiments with Spatial Restrictions at GeoCLEF 2007

The participation of the University of Hildesheim focused on the monolingual German and English tasks of GeoCLEF 2007. Based on the results of GeoCLEF 2005 and GeoCLEF 2006, the weighting and expansion of geographic Named Entities (NE) and Blind Relevance Feedback (BRF) were combined and an improved model for German Named Entity Recognition (NER) was evaluated. Post submission experiments are also presented.. A topic analysis revealed a wide spread of MAP values with high standard deviation values. Therefore further development will lie in the field of topic-adaptive systems.