An Award-Winning Tool
The best paper award for semantic web in-use and the best demo award at ESWC 2012 went to Explain-a-LOD!
Numerous statistics can be found, e.g., in the media. In many cases, those statistics consist only of a few attributes, in an extreme case, an identifier and a value. For example, the corruption perceptions index published by Transparency International relates a country to a numerical index depicting the perceived corruption in that country.
While finding a statistics is easy, finding interpretations that explain, e.g., why certain countries are more corrupt than others, is a more difficult task. There are tools for analyzing statistics, e.g., for finding correlations. However, such tools can only work well if the suspicious data is contained in the statistics file. For example, a country's GDP (gross domestic product) may be correlated with its corruption index. To discover this correlation with an analysis tool, the GDP needs to be included in the statistics. Especially when looking for explanations on a new problem, it is difficult to decide whether or not to include certain data in a statistics file.
Explain-a-LOD exploits Linked Open Data for generating additional attributes for a given statistics file. With the help of such dynamically created additional information, possible hypotheses for explaining the statistics are generated using correlation analysis and rule learning.
Talk about Explain-a-LOD at the Extended Semantic Web Conference (ESWC 2012):
This short video demonstrates how Explain-a-LOD works:
Note: Software is provided as is, without any warranty, including any damage done to hardware and/or software.
For any questions and/or suggestions concerning Explain-a-LOD, please contact Heiko Paulheim.