One of the most popular and trend-setting Internet applications is People Search on the World Wide Web. In its most general form, information extraction for persons from unstructured data is extremely challenging, and, we are pretty far away from satisfying solutions. However, current retrieval technology is able to cope with restricted variants of the problem, and this paper deals with such a variant, the so-called multi document person resolution. Given is a set of Web documents, and the task is to state for each document pair whether the two documents are talking about the same person or not.
For this problem Spock Inc., Silicon Valley, launched in 2007 a competition offering a grand prize of $50 000. Task was the person-specific classification of 100 000 Web pages within 4 hours on a standard PC, striving for a maximum F-Measure. The paper in hand describes the challenge and introduces the technology of the winning team from the Bauhaus University Weimar [see 1].
When studying heritage artefacts, and trying to represent what we know of them, it is important to portray not only key moments in their evolution, but also processes of transformation. In this contribution, we introduce a methodological framework of description of architectural changes, and investigate diagrammatic representations as means to visualize the above mentioned framework. We introduce two types of diagrams (diachrograms that distribute along a time axis transitions and states, variograms that detail the nature of the changes) that should help better understanding, how changes over time affect architecture. The paper also underlines key aspects of data in “historical sciences”: uncertainties, incompleteness, long ranges of time, unevenly distributed physical and temporal stratifications.
Companies are faced with managing as well as integrating large collections of distributed data today. Here, the challenging task is not to store these volumes of structured and interlinked data but to understand and analyze its explicit or implicit relationships. However, up to date there is virtually no support in navigating, visualizing or even analyzing structured data sets of this size appropriately. This paper describes novel rendering techniques enabling a new level of visual analytics combined with interactive exploration principles. The underlying visualization rationale is driven by the principle of providing detail information with respect to qualitative as well as quantitative aspects on user demand while offering an overview at any time. By means of our prototypical implementation and a real-world data set we show how to answer several data specific tasks by interactive visual exploration.
In the economic and financial analysis domain a quick access to the right information plays a major role. Using current systems, the search for and presentation of data is very cumbersome. The data, mostly in form of time-series, is stored in various databases. In order to retrieve the searched data, the analysts need to know where to search and sometimes even the structure of the database and its coding. Then it is required to export the data, process the data and create a chart to view the data. This might take time from tens of minutes to hours.
In our work we present a first prototype of an integrated search engine that takes as input a natural language query and offers graphic and text output depending on the user task. The system automatically identifies the time-series answers, types of graphical data presentation and shows the results in a web browser and in Excel. The knowledge-based expert system uses domain ontologies for extraction of economic terms in the search queries and specially built data type taxonomy with user task and chart type ontologies for identification of graphic output.
An increasing amount of valuable information is stored in RDF. In order to let humans access this information, providing an appropriate visualization of RDF data is an important challenge. In this paper, we present a new approach, combining list and a graph visualization to counterbalance the respective disadvantages of both representation paradigms to better handle the complexity of both the size and the
structure of RDF data.
In recent years the visualization of knowledge has been gaining wider attention: visualization is said to enhance human capabilities for knowledge intense activities such as decision making and strategic thinking. However, this is a recent field and still widely unexplored. Thus far, the advantages of knowledge visualization have been investigated mainly through anecdotal evidence and qualitative studies. In this paper, we propose an experimental approach to further comprehend the role of visualization in fostering knowledge sharing. We plan to compare the elicitation and evaluation processes of groups who are provided (1) with an optimal visual support, (2) with a sub-optimal visual support, and (3) without any visualization. The goal of our research is to apply the experimental approach – widely used in studying GSS (Group Support System) but seldom used in knowledge management – to shed light on the role of visualization for knowledge-intensive tasks in groups. We report first preliminary results of an experiment with 56 MBA students and also outline the limitations of our approach.