@inproceedings{giannakidou2008coclustering, abstract = {Under social tagging systems, a typical Web 2.0 application, users label digital data sources by using freely chosen textual descriptions (tags). Poor retrieval in the aforementioned systems remains a major problem mostly due to questionable tag validity and tag ambiguity. Earlier clustering techniques have shown limited improvements, since they were based mostly on tag co-occurrences. In this paper, a co-clustering approach is employed, that exploits joint groups of related tags and social data sources, in which both social and semantic aspects of tags are considered simultaneously. Experimental results demonstrate the efficiency and the beneficial outcome of the proposed approach in correlating relevant tags and resources.}, author = {Giannakidou, Eirini and Koutsonikola, Vassiliki A. and Vakali, Athena and Kompatsiaris, Yiannis}, booktitle = {WAIM}, crossref = {conf/waim/2008}, ee = {http://dx.doi.org/10.1109/WAIM.2008.61}, file = {giannakidou2008coclustering.pdf:giannakidou2008coclustering.pdf:PDF}, groups = {public}, interhash = {bf55ee73fa8e8e370cffe8ef7bb9cd60}, intrahash = {2b24046689df977f7853b557c04689f3}, isbn = {978-0-7695-3185-4}, pages = {317-324}, publisher = {IEEE}, timestamp = {2011-02-17 11:00:40}, title = {Co-Clustering Tags and Social Data Sources.}, url = {http://dblp.uni-trier.de/db/conf/waim/waim2008.html#GiannakidouKVK08}, username = {dbenz}, year = 2008 } @article{cilibrasi2005clustering, abstract = { We present a new method for clustering based on compression. The method does not use subject-specific features or background knowledge, and works as follows: First, we determine a parameter-free, universal, similarity distance, the normalized compression distance or NCD, computed from the lengths of compressed data files (singly and in pairwise concatenation). Second, we apply a hierarchical clustering method. The NCD is not restricted to a specific application area, and works across application area boundaries. A theoretical precursor, the normalized information distance, co-developed by one of the authors, is provably optimal. However, the optimality comes at the price of using the noncomputable notion of Kolmogorov complexity. We propose axioms to capture the real-world setting, and show that the NCD approximates optimality. To extract a hierarchy of clusters from the distance matrix, we determine a dendrogram (ternary tree) by a new quartet method and a fast heuristic to implement it. The method is implemented and available as public software, and is robust under choice of different compressors. To substantiate our claims of universality and robustness, we report evidence of successful application in areas as diverse as genomics, virology, languages, literature, music, handwritten digits, astronomy, and combinations of objects from completely different domains, using statistical, dictionary, and block sorting compressors. In genomics, we presented new evidence for major questions in Mammalian evolution, based on whole-mitochondrial genomic analysis: the Eutherian orders and the Marsupionta hypothesis against the Theria hypothesis.}, author = {Cilibrasi, R. and Vitanyi, P.M.B.}, doi = {10.1109/TIT.2005.844059}, interhash = {2016d3da3ebb9d17fdf0be152c2f2069}, intrahash = {5156d51daa332b82b27cc4665dbff1f5}, issn = {0018-9448}, journal = {IEEE Transactions on Information Theory}, month = {April}, number = 4, pages = { 1523-1545}, title = {Clustering by compression}, volume = 51, year = 2005 } @article{talavera2001generalitybased, address = {Los Alamitos, CA, USA}, author = {Talavera, Luis and B{\'e}jar, Javier}, doi = {10.1109/34.908969,}, interhash = {c6c47f26f4793b3fedd46209796e792c}, intrahash = {358a8cba2b3442748874b422fb28e7f9}, issn = {0162-8828}, journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence}, number = 2, pages = {196-206}, publisher = {IEEE Computer Society}, title = {Generality-Based Conceptual Clustering with Probabilistic Concepts}, url = {http://www.computer.org/portal/web/csdl/doi/10.1109/34.908969}, volume = 23, year = 2001 } @inproceedings{vinh2009information, abstract = {Information theoretic based measures form a fundamental class of similarity measures for comparing clusterings, beside the class of pair-counting based and set-matching based measures. In this paper, we discuss the necessity of correction for chance for information theoretic based measures for clusterings comparison. We observe that the baseline for such measures, i.e. average value between random partitions of a data set, does not take on a constant value, and tends to have larger variation when the ratio between the number of data points and the number of clusters is small. This effect is similar in some other non-information theoretic based measures such as the well-known Rand Index. Assuming a hypergeometric model of randomness, we derive the analytical formula for the expected mutual information value between a pair of clusterings, and then propose the adjusted version for several popular information theoretic based measures. Some examples are given to demonstrate the need and usefulness of the adjusted measures.}, address = {New York, NY, USA}, author = {Vinh, Nguyen Xuan and Epps, Julien and Bailey, James}, booktitle = {ICML '09: Proceedings of the 26th Annual International Conference on Machine Learning}, doi = {10.1145/1553374.1553511}, interhash = {ddd96b934438029873242aeabc26a201}, intrahash = {bed9702898bc8c50faa21eabd068b8d9}, isbn = {978-1-60558-516-1}, location = {Montreal, Quebec, Canada}, pages = {1073--1080}, publisher = {ACM}, title = {Information theoretic measures for clusterings comparison: is a correction for chance necessary?}, url = {http://portal.acm.org/citation.cfm?id=1553511}, year = 2009 } @inproceedings{shepitsen2008personalized, abstract = {Collaborative tagging applications allow Internet users to annotate resources with personalized tags. The complex network created by many annotations, often called a folksonomy, permits users the freedom to explore tags, resources or even other user's profiles unbound from a rigid predefined conceptual hierarchy. However, the freedom afforded users comes at a cost: an uncontrolled vocabulary can result in tag redundancy and ambiguity hindering navigation. Data mining techniques, such as clustering, provide a means to remedy these problems by identifying trends and reducing noise. Tag clusters can also be used as the basis for effective personalized recommendation assisting users in navigation. We present a personalization algorithm for recommendation in folksonomies which relies on hierarchical tag clusters. Our basic recommendation framework is independent of the clustering method, but we use a context-dependent variant of hierarchical agglomerative clustering which takes into account the user's current navigation context in cluster selection. We present extensive experimental results on two real world dataset. While the personalization algorithm is successful in both cases, our results suggest that folksonomies encompassing only one topic domain, rather than many topics, present an easier target for recommendation, perhaps because they are more focused and often less sparse. Furthermore, context dependent cluster selection, an integral step in our personalization algorithm, demonstrates more utility for recommendation in multi-topic folksonomies than in single-topic folksonomies. This observation suggests that topic selection is an important strategy for recommendation in multi-topic folksonomies.}, address = {New York, NY, USA}, author = {Shepitsen, Andriy and Gemmell, Jonathan and Mobasher, Bamshad and Burke, Robin}, booktitle = {RecSys '08: Proceedings of the 2008 ACM conference on Recommender systems}, doi = {http://doi.acm.org/10.1145/1454008.1454048}, interhash = {c9028129dd7cd8314673bd64cbb6198e}, intrahash = {a7552f8d8d5db4f867ae6e94e1a4442f}, isbn = {978-1-60558-093-7}, location = {Lausanne, Switzerland}, pages = {259--266}, publisher = {ACM}, title = {Personalized recommendation in social tagging systems using hierarchical clustering}, url = {http://portal.acm.org/citation.cfm?id=1454008.1454048}, year = 2008 } @inproceedings{delling2007engineering, author = {Delling, Daniel and Gaertler, Marco and G{\"o}rke, Robert and Wagner, Dorothea}, booktitle = {Proceedings of the European Conference of Complex Systems (ECCS'07)}, interhash = {b0b92b2ead46ef60435173a6fb803045}, intrahash = {48417fa551e51439159e5fdd575825df}, month = {October}, note = {as poster}, pdf = {http://i11www.ira.uka.de/algo/people/rgoerke/publications/pdf/dggw-ecgc-07_poster.pdf}, title = {Engineering Comparators for Graph Clusterings}, url = {http://i11www.ira.uka.de/algo/people/rgoerke/publications/pdf/dggw-ecgc-07_poster.pdf}, year = 2007 } @inproceedings{delling2007engineeringa, author = {Delling, Daniel and Gaertler, Marco and G{\"o}rke, Robert and Nikoloski, Zoran and Wagner, Dorothea}, booktitle = {Proceedings of the European Conference of Complex Systems (ECCS'07)}, interhash = {5c88c6ad4f9a66de094125b3ce600a55}, intrahash = {fa6b1f4966b69da84f9582c2aba82cab}, month = {October}, note = {as poster}, pdf = {http://i11www.ira.uka.de/algo/people/rgoerke/publications/pdf/dggnw-eect-07_poster.pdf}, title = {Engineering the Evaluation of Clustering Techniques}, url = {http://i11www.ira.uka.de/algo/people/rgoerke/publications/pdf/dggnw-eect-07_poster.pdf}, year = 2007 } @inproceedings{cattuto2007emergent, address = {Dresden, Germany}, author = {Cattuto, Ciro and Baldassarri, Andrea and Servedio, Vito D. P. and Loreto, Vittorio}, booktitle = {Proceedings of the European Confeence on Complex Systems}, interhash = {9afde66e2d53e2f23bed303f7bda30af}, intrahash = {3977cdaf1ce7a4c500ac5cfd5a91c9e5}, month = {October}, title = {Emergent Community Structure in Social Tagging Systems}, year = 2007 } @misc{capocci2007taxonomy, abstract = { In this paper we investigate the nature and structure of the relation between imposed classifications and real clustering in a particular case of a scale-free network given by the on-line encyclopedia Wikipedia. We find a statistical similarity in the distributions of community sizes both by using the top-down approach of the categories division present in the archive and in the bottom-up procedure of community detection given by an algorithm based on the spectral properties of the graph. Regardless the statistically similar behaviour the two methods provide a rather different division of the articles, thereby signaling that the nature and presence of power laws is a general feature for these systems and cannot be used as a benchmark to evaluate the suitability of a clustering method.}, author = {Capocci, A. and Rao, F. and Caldarelli, G.}, interhash = {df8a20aa40cce46aa0adf4f6360664dc}, intrahash = {9c69bc97d22b7e5c2d90d8765b491a16}, title = {Taxonomy and clustering in collaborative systems: the case of the on-line encyclopedia Wikipedia}, url = {http://www.citebase.org/abstract?id=oai:arXiv.org:0710.3058}, year = 2007 } @misc{newman2003structure, abstract = {Inspired by empirical studies of networked systems such as the Internet, social networks, and biological networks, researchers have in recent years developed a variety of techniques and models to help us understand or predict the behavior of these systems. Here we review developments in this field, including such concepts as the small-world effect, degree distributions, clustering, network correlations, random graph models, models of network growth and preferential attachment, and dynamical processes taking place on networks.}, author = {Newman, M. E. J.}, file = {newman2003structure.pdf:newman2003structure.pdf:PDF}, interhash = {7bedd01cb4c06af9f5200b0fb3faa571}, intrahash = {d53568209eef08fb0a8734cf34c59a71}, lastdatemodified = {2006-10-07}, lastname = {Newman}, month = {March}, own = {notown}, pdf = {newman03-structure.pdf}, read = {notread}, title = {The structure and function of complex networks}, url = {http://arxiv.org/abs/cond-mat/0303516}, year = 2003 } @book{jain1988algorithms, address = {Upper Saddle River, NJ, USA}, author = {Jain, Anil K. and Dubes, Richard C.}, file = {jain1988algorithms.pdf:jain1988algorithms.pdf:PDF}, interhash = {443a79c152c5681cdc664714b50d116c}, intrahash = {4a1adbfdc7b83b201dd8fb3e5a109609}, lastdatemodified = {2007-03-13}, lastname = {Jain}, note = {Attention: PDF is rather large (~39MB)}, own = {notown}, pdf = {jain88_algorithms.pdf}, publisher = {Prentice-Hall, Inc.}, read = {notread}, title = {Algorithms for clustering data}, url = {http://portal.acm.org/citation.cfm?id=46712}, year = 1988 } @inproceedings{brooks2006improved, abstract = {Tags have recently become popular as a means of annotating and organizing Web pages and blog entries. Advocates of tagging argue that the use of tags produces a 'folksonomy', a system in which the meaning of a tag is determined by its use among the community as a whole. We analyze the effectiveness of tags for classifying blog entries by gathering the top 350 tags from Technorati and measuring the similarity of all articles that share a tag. We find that tags are useful for grouping articles into broad categories, but less effective in indicating the particular content of an article. We then show that automatically extracting words deemed to be highly relevant can produce a more focused categorization of articles. We also show that clustering algorithms can be used to reconstruct a topical hierarchy among tags, and suggest that these approaches may be used to address some of the weaknesses in current tagging systems.}, address = {New York, NY, USA}, author = {Brooks, Christopher H. and Montanez, Nancy}, booktitle = {WWW '06: Proceedings of the 15th international conference on World Wide Web}, file = {:brooks06-improved.pdf:PDF;brooks2006improved.pdf:brooks2006improved.pdf:PDF}, groups = {public}, interhash = {c88a665abf8d88c5a7ae95fa2783f837}, intrahash = {5c9c83e89da2faa8906a5927fe7ca3ef}, lastdatemodified = {2006-07-18}, lastname = {Brooks}, longnotes = {[[http://www2006.org/programme/files/pdf/583-slides.pdf slides]] Summary: - authors analyse the effectiveness of tags for classifying blog articles (technorati) - clustering of articles beloning to top 350 technorati tags * by tag * randomly * by related by Google News - results: * tags help to classify articles into broad categories (yet Google News performs better) * tags are not that descriptive for a specific topic of an article * automatically extracted tags (by TF/IDF) are much more descriptive for specific content - 2nd study: hierarchical clustering of articles (starting from tag clusters, i.e. all articles who share a tag) - resulting tag hierarchy comes close to e.g. Yahoo hand-built one}, own = {own}, pages = {625--632}, pdf = {brooks06-improved.pdf}, publisher = {ACM Press}, read = {read}, timestamp = {2009-09-29 16:23:07}, title = {Improved annotation of the blogosphere via autotagging and hierarchical clustering}, url = {http://www2006.org/programme/item.php?id=583}, username = {dbenz}, year = 2006 } @inproceedings{ramage2009clustering, abstract = {Automatically clustering web pages into semantic groups promises improved search and browsing on the web. In this paper, we demonstrate how user-generated tags from largescale social bookmarking websites such as del.icio.us can be used as a complementary data source to page text and anchor text for improving automatic clustering of web pages. This paper explores the use of tags in 1) K-means clustering in an extended vector space model that includes tags as well as page text and 2) a novel generative clustering algorithm based on latent Dirichlet allocation that jointly models text and tags. We evaluate the models by comparing their output to an established web directory. We find that the naive inclusion of tagging data improves cluster quality versus page text alone, but a more principled inclusion can substantially improve the quality of all models with a statistically significant absolute F-score increase of 4%. The generative model outperforms K-means with another 8% F-score increase.}, address = {New York, NY, USA}, author = {Ramage, Daniel and Heymann, Paul and Manning, Christopher D. and Garcia-Molina, Hector}, booktitle = {WSDM '09: Proceedings of the Second ACM International Conference on Web Search and Data Mining}, doi = {http://doi.acm.org/10.1145/1498759.1498809}, file = {ramage2009clustering.pdf:ramage2009clustering.pdf:PDF}, groups = {public}, interhash = {5595f06f88310ed67fd6fe23f813c69b}, intrahash = {75c4bad29d7eb4b34f68da27f0353516}, isbn = {978-1-60558-390-7}, location = {Barcelona, Spain}, pages = {54--63}, publisher = {ACM}, timestamp = {2009-04-24 10:19:45}, title = {Clustering the tagged web}, url = {http://portal.acm.org/citation.cfm?id=1498809}, username = {dbenz}, year = 2009 } @inproceedings{cimiano2004comparing, abstract = {The application of clustering methods for automatic taxonomy construction from text requires knowledge about the tradeoff between, (i), their effectiveness (quality of result), (ii), efficiency (run-time behaviour), and, (iii), traceability of the taxonomy construction by the ontology engineer. In this line, we present an original conceptual clustering method based on Formal Concept Analysis for automatic taxonomy construction and compare it with hierarchical agglomerative clustering and hierarchical divisive clustering.}, author = {Cimiano, Philipp and Hotho, Andreas and Staab, Steffen}, booktitle = {ECAI 2004 Proceedings of the 16th European Conference on Artificial Intelligence, 22 - 27 August, Valencia, Spain}, editor = {de M\'{a}ntaras, R. L\'{o}pez and Saitta, L.}, file = {cimiano2004comparing.pdf:cimiano2004comparing.pdf:PDF}, groups = {public}, interhash = {5ebc73142f0c4d51a1037432435bab94}, intrahash = {4e2f4ba3e051f120c2bc8216aad7cdaa}, pages = {435-439}, publisher = {IOS Press}, timestamp = {2011-02-02 13:38:11}, title = {Comparing Conceptual, Divise and Agglomerative Clustering for Learning Taxonomies from Text}, username = {dbenz}, year = 2004 } @inproceedings{grineva2008harnessing, abstract = {The quality of the current tagging services can be greatly improved if the service is able to cluster tags by their meaning. Tag clouds clustered by higher level topics enable the users to explore their tag space, which is especially needed when tag clouds become large. We demonstrate TagCluster - a tool for automated tag clustering that harnesses knowledge from Wikipedia about semantic relatedness between tags and names of categories to achieve smart clustering. Our approach shows much better quality of clusters compared to the existing techniques that rely on tag co-occurrence analysis in the tagging service.}, author = {Grineva, Maria and Grinev, Maxim and Turdakov, Denis and Velikhov, Pavel}, booktitle = {Proceedings of the International Workshop on Knowledge Acquisition from the Social Web (KASW2008)}, file = {grineva2008harnessing.pdf:grineva2008harnessing.pdf:PDF}, groups = {public}, interhash = {814ebc26a00c8facc9d2a7ef3edd256e}, intrahash = {093e8262f1cf4f2c4a159b5d7b76ce78}, timestamp = {2011-02-02 14:57:13}, title = {Harnessing Wikipedia for Smart Tags Clustering}, username = {dbenz}, year = 2008 } @article{Luo20091271, abstract = {Clustering is a very powerful data mining technique for topic discovery from text documents. The partitional clustering algorithms, such as the family of k-means, are reported performing well on document clustering. They treat the clustering problem as an optimization process of grouping documents into k clusters so that a particular criterion function is minimized or maximized. Usually, the cosine function is used to measure the similarity between two documents in the criterion function, but it may not work well when the clusters are not well separated. To solve this problem, we applied the concepts of neighbors and link, introduced in [S. Guha, R. Rastogi, K. Shim, ROCK: a robust clustering algorithm for categorical attributes, Information Systems 25 (5) (2000) 345-366], to document clustering. If two documents are similar enough, they are considered as neighbors of each other. And the link between two documents represents the number of their common neighbors. Instead of just considering the pairwise similarity, the neighbors and link involve the global information into the measurement of the closeness of two documents. In this paper, we propose to use the neighbors and link for the family of k-means algorithms in three aspects: a new method to select initial cluster centroids based on the ranks of candidate documents; a new similarity measure which uses a combination of the cosine and link functions; and a new heuristic function for selecting a cluster to split based on the neighbors of the cluster centroids. Our experimental results on real-life data sets demonstrated that our proposed methods can significantly improve the performance of document clustering in terms of accuracy without increasing the execution time much.}, author = {Luo, Congnan and Li, Yanjun and Chung, Soon M.}, doi = {10.1016/j.datak.2009.06.007}, interhash = {bf59c4cf26cbc35d6142630b34a66d37}, intrahash = {13483e90d8b46ef9435ec71473aacee4}, issn = {0169-023X}, journal = {Data & Knowledge Engineering}, note = {Including Special Section: Conference on Privacy in Statistical Databases (PSD 2008) - Six selected and extended papers on Database Privacy}, number = 11, pages = {1271 - 1288}, title = {Text document clustering based on neighbors}, url = {http://www.sciencedirect.com/science/article/B6TYX-4WNB4Y8-1/2/1dcd00d9c049988da53b44a526dd6555}, volume = 68, year = 2009 } @misc{Farrahi_discoveringhuman, abstract = {We present a framework to automatically discover people’s routines from information extracted by cell phones. The framework is built from a probabilistic topic model learned on novel bag type representations of activity-related cues (location, proximity and their temporal variations over a day) of peoples ’ daily routines. Using real-life data from the Reality Mining dataset, covering 68 000+ hours of human activities, we can successfully discover location-driven (from cell tower connections) and proximity-driven (from Bluetooth information) routines in an unsupervised manner. The resulting topics meaningfully characterize some of the underlying co-occurrence structure of the activities in the dataset, including “going to work early/late”, “being home all day”, “working constantly”, “working sporadically” and “meeting at lunch time”. 1.}, author = {Farrahi, Katayoun and Gatica-perez, Daniel}, interhash = {5e3f9c64f6fb9ba5226e3345acd3ddd8}, intrahash = {4c905f2cfc5e88c271ebc4f10d47de30}, title = {Discovering Human Routines from Cell Phone Data with Topic Models}, url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.139.5105}, year = 2010 } @inproceedings{lu2009exploit, abstract = {In this poster, we investigate how to enhance web clustering by leveraging the tripartite network of social tagging systems. We propose a clustering method, called "Tripartite Clustering", which cluster the three types of nodes (resources, users and tags) simultaneously based on the links in the social tagging network. The proposed method is experimented on a real-world social tagging dataset sampled from del.icio.us. We also compare the proposed clustering approach with K-means. All the clustering results are evaluated against a human-maintained web directory. The experimental results show that Tripartite Clustering significantly outperforms the content-based K-means approach and achieves performance close to that of social annotation-based K-means whereas generating much more useful information.}, address = {New York, NY, USA}, author = {Lu, Caimei and Chen, Xin and Park, E. K.}, booktitle = {CIKM '09: Proceeding of the 18th ACM conference on Information and knowledge management}, doi = {10.1145/1645953.1646167}, interhash = {e192e53972f28d78f1ecbffbfea08bed}, intrahash = {a120cece36e15b12321c87e7d0938d73}, isbn = {978-1-60558-512-3}, location = {Hong Kong, China}, pages = {1545--1548}, publisher = {ACM}, title = {Exploit the tripartite network of social tagging for web clustering}, url = {http://portal.acm.org/citation.cfm?id=1646167&dl=GUIDE&coll=GUIDE&CFID=93888742&CFTOKEN=72927742}, year = 2009 } @misc{Karypis02multilevelhypergraph, abstract = {Introduction Hypergraph partitioning is an important problem with extensive application to many areas, including VLSI design [Alpert and Kahng, 1995], efficient storage of large databases on disks [Shekhar and Liu, 1996], and data mining [Mobasher et al., 1996, Karypis et al., 1999b]. The problem is to partition the vertices of a hypergraph into k equal-size parts, such that the number of hyperedges connecting vertices in different parts is minimized. During the course of VLSI circuit design and synthesis, it is important to be able to divide the system specification into clusters so that the inter-cluster connections are minimized. This step has many applications including design packaging, HDL-based synthesis, design optimization, rapid prototyping, simulation, and testing. Many rapid prototyping systems use partitioning to map a complex circuit onto hundreds of interconnected FPGAs. Such partitioning instances are challenging because the timing, area, and I/O resource utilization }, author = {Karypis, George}, interhash = {c79f1aad4b40640a346bd67fdd4eada3}, intrahash = {e1d8b31de59731bbf41a8559c8cf9caa}, title = {Multilevel Hypergraph Partitioning}, url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.6.9117}, year = 2002 } @inproceedings{grahl2007clustering, abstract = {Currently, social bookmarking systems provide intuitive support for browsing locally their content. A global view is usually presented by the tag cloud of the system, but it does not allow a conceptual drill-down, e. g., along a conceptual hierarchy. In this paper, we present a clustering approach for computing such a conceptual hierarchy for a given folksonomy. The hierarchy is complemented with ranked lists of users and resources most related to each cluster. The rankings are computed using our FolkRank algorithm. We have evaluated our approach on large scale data from the del.icio.us bookmarking system.}, address = {Graz, Austria}, author = {Grahl, Miranda and Hotho, Andreas and Stumme, Gerd}, booktitle = {7th International Conference on Knowledge Management (I-KNOW '07)}, interhash = {5cf58d2fdd3c17f0b0c54ce098ff5b60}, intrahash = {334d3ab11400c4a3ea3ed5b1e95c1855}, issn = {0948-695x}, month = sep, pages = {356-364}, publisher = {Know-Center}, title = {Conceptual Clustering of Social Bookmarking Sites}, url = {/brokenurl#www.tagora-project.eu/wp-content/2007/06/grahl_iknow07.pdf}, vgwort = {14}, year = 2007 }