Zeljko Carevic

Thomas Krämer, Zeljko Carevic, Dwaipayan Roy et al.

In this demo paper, we present ConSTR, a novel Contextual Search Term Recommender that utilises the user's interaction context for search term recommendation and literature retrieval. ConSTR integrates a two-layered recommendation interface: the first layer suggests terms with respect to a user's current search term, and the second layer suggests terms based on the users' previous search activities (interaction context). For the demonstration, ConSTR is built on the arXiv, an academic repository consisting of 1.8 million documents.

Joeran Beel, Zeljko Carevic, Johann Schaible et al.

Recommender-system datasets are used for recommender-system evaluations, training machine-learning algorithms, and exploring user behavior. While there are many datasets for recommender systems in the domains of movies, books, and music, there are rather few datasets from research-paper recommender systems. In this paper, we introduce RARD, the Related-Article Recommendation Dataset, from the digital library Sowiport and the recommendation-as-a-service provider Mr. DLib. The dataset contains information about 57.4 million recommendations that were displayed to the users of Sowiport. Information includes details on which recommendation approaches were used (e.g. content-based filtering, stereotype, most popular), what types of features were used in content based filtering (simple terms vs. keyphrases), where the features were extracted from (title or abstract), and the time when recommendations were delivered and clicked. In addition, the dataset contains an implicit item-item rating matrix that was created based on the recommendation click logs. RARD enables researchers to train machine learning algorithms for research-paper recommendations, perform offline evaluations, and do research on data from Mr. DLib's recommender system, without implementing a recommender system themselves. In the field of scientific recommender systems, our dataset is unique. To the best of our knowledge, there is no dataset with more (implicit) ratings available, and that many variations of recommendation algorithms. The dataset is available at http://data.mr-dlib.org, and published under the Creative Commons Attribution 3.0 Unported (CC-BY) license.

Zeljko Carevic, Thomas Krichel, Philipp Mayr

In this paper, we present an approach for analyzing the behavior of editors in the large current awareness service "NEP: New Economics Papers". We processed data from more than 38,000 issues derived from 90 different NEP reports over the past ten years. The aim of our analysis was to gain an inside to the editor behaviour when creating an issue and to look for factors that influence the success of a report. In our study we looked at the following features: average editing time, the average number of papers in an issue and the editor effort measured on presorted issues as relative search length (RSL). We found an average issue size of 12.4 documents per issue. The average editing time is rather low with 14.5 minute. We get to the point that the success of a report is mainly driven by its topic and the number of subscribers, as well as proactive action by the editor to promote the report in her community.

Zeljko Carevic, Philipp Mayr

In digital libraries recommendations can be valuable for researchers, e.g. recommending related literature to a given context. Typically, in a scientific context the simple presentation of related content is not sufficient. Often the users demand a more detailed view on the connection of a document and its specific recommendations. The aim of pennants introduced by Howard White (2007) is to provide the user with a graph showing the relatedness / distance between a given document and related documents. Co-citation but also co-occurrence analysis are established methods for finding related documents to a seed. A seed could be for instance an author, a keyword, or a publication. In this paper we introduce a recommender system in the digital library sowiport using pennant diagrams which can be created from co-citation and/or co-occurrence analysis. The presentation at the NKOS workshop will present demos of pennants in sowiport and will elaborate on practical questions in visualizing pennants and evaluating the utility of pennants for search.