K.U.Leuven University Library (Belgium) developed an experimental interface for subject search by UDC in the OPAC. The interface combines the search facilities of a classification with those of a word system, since it enables the end user to search by subject terms and to see these terms in the hierarchy of broader, parallel and more specific terms. This project should be seen as an important indication of the libraryâ s growing concern to present its information sources in a content-structured and user-friendly way. At the same time, it has to be situated in a new policy for knowledge organization, which aims to find a balance between the local and overall needs of a library network. Finally, this project comes at a moment when K.U.Leuven University Library is in full conversion to Aleph 500 software.

The presentation was part of The 6th European Networked Knowledge Organization Systems (NKOS) Workshop at the 11th European Conference on Research and Advanced Technology for Digital Libraries (ECDL), Budapest, Hungary September 21, 2007 (http://www.comp.glam.ac.uk/pages/research/hypermedia/nkos/nkos2007/programme.html) / The presentation will briefly introduce a series of major principles for bringing subject terminology to the network level. A closer look at one KOS in particular, the Dewey Decimal Classification, should help to gain more insight into the perceived difficulties and potential benefits of building taxonomy services out and on top of classic large-scale vocabularies or taxonomies.

This is a 27-slide presentation made at ISKO 2004 (International Society for Knowledge Organization) in London, UK. Recent projects in subject indexing and classification at K.U.Leuven University Library (Belgium) aim to give new impulses to knowledge organization within the institution. While in recent years a lot of attention was given, and with good reason, to the technical and administrative integration of e-sources, less energy was invested in organizing the content of traditional and electronic collections. Nevertheless, presenting information sources in a content-structured way remains a core task of our University Library. This paper focuses on some experiments with subject search interfaces at K.U.Leuven University Library and situates them in a new policy for knowledge organization, which tries to find a balance between local and overall needs.

One of the major tasks in Data Mining is classification. The growing of Decision Tree from data is a very efficient technique for learning classifiers. The selection of an attribute used to split the data set at each Decision Tree node is fundamental to properly classify objects; a good selection will improve the accuracy of the classification. In this paper, we study the behavior of the Decision Trees induced with 14 attribute selection measures over three data sets taken from UCI Machine Learning Repository.

Social classification is the process in which a community of users categorizes the resources in that community for their own use. Given enough users and categorization, this will lead to any given resource being represented by a set of labels or descriptors shared throughout the community (Mathes, 2004). Social classification has become an extremely popular way of structuring online communities in recent years. Well-known examples of such communities are the bookmarking websites Furl (http://www.furl.net/) and del.icio.us (http://del.icio.us/), and Flickr (http://www.flickr.com/) where users can post their own photos and tag them. Social classification, however, is not limited to tagging resources: another possibility is to tag people, examples of which are Consumating (http://www.consumating.com/), a collaborative tag-based personals website, and Kevo (http://www.kevo.com/), a website that lets users tag and contribute media and information on celebrities. Another application of people tagging is expertise classification, an emerging subfield of social classification. Here, members of a group or community are classified and ranked based on the expertise they possess on a particular topic. Expertise classification is essentially comprised of two different components: expertise tagging and expert ranking. Expertise tagging focuses on describing one person at a time by assigning tags that capture that person's topical expertise, such as â speech recognition' or â small-world networks'. information request, such as, for instance, a query submitted to a search engine. Methods are developed to combine the information about individual members' expertise (tags), to provide on-the-fly query-driven rankings of community members. Expertise classification can be done in two principal ways. The simplest option follows the principle of social bookmarking websites: members are asked to supply tags that describe their own expertise and to rank the other community members with regard to a specific request for information. Alternatively, automatic expertise classification ideally extracts expertise terms automatically from a user's documents and e-mails by looking for terms that are representative for that user. These terms are then matched on the information request to produce an expert ranking of all community members. In this paper we describe such an automatic method of expertise classification and evaluate it using human expertise classification judgments. In the next section we will describe some of the related work on expertise classification, after which we will describe our automatic method of expertise classification and our evaluation of them in sections 3 and 4. Sections 5.1 and 5.1 describe our findings on expertise tagging and expert rankings, followed by discussion and our conclusions in section 6 and recommendations for future work in section 7.

Artificial Intelligence Lab, Department of MIS, University of Arizona / The problems of information overload and vocabulary differences have become more pressing with the emergence of the increasingly popular Internet services. The main information retrieval mechanisms provided by the prevailing Internet WWW software are based on either keyword search or hypertext browsing. Keyword search often results in low precision, poor recall, and slow response time due to the limitations of indexing and communication methods, controlled language based interfaces, and the inability of searchers themselves to articulate their needs fully. Hypertext browsing, on the other hand, allows users to explore only a very small portion of a large Internet information space. A large information space can also potentially confuse and disorient its user and it can cause the user to spend a great deal of time while learning nothing specific. This research aims to provide concept-based categorization and search capabilities for Internet WWW servers based on selected machine learning and parallel computing techniques. Our proposed approach, which is grounded on automatic textual analysis of Internet documents, attempts to address the Internet search problem by first categorizing the content of Internet documents and subsequently providing semantic search capabilities based on a concept space approach. As a first step, we propose a multi-layered neural network clustering algorithm employing the Kohonen self-organizing feature map to categorize the Internet homepages according to their content. The category hierarchies created could serve to partition the vast Internet services into subject-specific categories and databases. After individual subject categories have been created, we propose to generate domain-specific concept spaces for each subject category. The concept spaces can then be used to support concept-based information retrieval, a significant improvement over the existing keyword searching and hypertext browsing options for Internet resource discovery. As Internet information space continues to grow at the present pace, we believe this research would shed light on potentially robust and scalable solutions to the increasingly complex and urgent information access and sharing problems that are certain to emerge in the future Internet society.

There is increasing evidence that libraries, traditional and digital, must support learning, especially the acquisition and enhancement of scientific reasoning skills. This paper discusses how classificatory structures, such as a faceted thesaurus, can be enhancedfor novice science learning. Physical geography is used as the domain discipline, and the Alexandria Digital Earth Prototype project provides the test bed for instructional materials and user analyses. The use of concept maps and topic maps for developing digital learning spaces is briefly discussed.

No description available.

This project explored the strategies that undergraduate and graduate chemistry students engaged in when solving classification tasks involving microscopic (particulate) representations of chemical substances and microscopic and symbolic representations of different chemical reactions. We were specifically interested in characterizing the basic features to which students pay attention while classifying, identifying the patterns of reasoning that they follow, and comparing the performance of students with different levels of preparation in the discipline. In general, our results suggest that advanced levels of expertise in chemical classification do not necessarily evolve in a linear and continuous way with academic training. Novice students had a tendency to reduce the cognitive demand of the task and rely on common-sense reasoning; they had difficulties differentiating concepts (conceptual undifferentiation) and based their classification decisions on only one variable (reduction). These ways of thinking lead them to consider extraneous features, pay more attention to explicit or surface features than implicit features and to overlook important and relevant features. However, unfamiliar levels of representations (microscopic level) seemed to trigger deeper and more meaningful thinking processes. On the other hand, expert students classified entities using a specific set of rules that they applied throughout the classification tasks. They considered a larger variety of implicit features and the unfamiliarity with the microscopic level of representation did not affect their reasoning processes. Consequently, novices created numerous small groups, few of them being chemically meaningful, while experts created few but large chemically meaningful groups. Novices also had difficulties correctly classifying entities in chemically meaningful groups. Finally, expert chemists in our study used classification schemes that are not necessarily traditionally taught in classroom chemistry (e.g. the structure of substances is more relevant to them than their composition when classifying substances as compounds or elements). This result suggests that practice in the field may develop different types of knowledge framework than those usually presented in chemistry textbooks.

The procolophonid amniote Barasaurus besairiei Piveteau 1955 is fully described and restored for the first time with emphasis placed on the postcranial skeleton, which is only poorly known in most of the other taxa of early amniotes. / The study focuses on testing a hypothesis of relationships, namely whether procolophonids are the sister-group of Testudines as proposed by Reisz & Laurin (1991). The description provides a sound basis for a new phylogenetic study of early amniotes. Using 13 taxa and 68 characters, the analysis indicates that synapsids are the sister-group of all other known amniotes, named Sauropsida. The Sauropsida are divided into Palaeosauropsida and Eusauropsida. Palaeosauropsida comprise Millerettidae as the sister-group of Procolophoniformes. The Procolophoniformes contain Procolophonia and Testudinomorpha as sister-groups. Testudines are the sister-group of Pareiasauria within the Testudinomorpha. Within Procolophonia, the family Owenettidae, including Barasaurus and Owenetta, is the sister-group of the family Procolophonidae. Eusauropsida include captorhinids, Palaeothyris and diapsids. / All of the three major amniote clades have extant taxa: Synapsida--mammals; Palaeosauropsida--turtles; Eusauropsida--diapsids including birds. The terms "Reptilia" and "Parareptilia" are omitted from systematics: Parareptilia for a misleading name and Reptilia in general because of its historical burden. / The new tree is strong in supporting Procolophonia and Testudinomorpha (sister-group of Pareiasauria and Testudines). It is not very firm in establishing eusauropsids and diadectomorphs because they were outside the main focus of the analysis. Mesosauria is the only group of Palaeozoic amniotes not included in this study.