The ongoing transition from the current Web of unstructured data to the Data Web yet requires scalable and accurate approaches for the extraction of structured data in RDF (Resource Description Framework). One of the key steps towards extracting RDF from natural-language corpora is the disambiguation of named entities.
AGDISTIS combines the HITS algorithm with label expansion strategies and string similarity measures. Based on this combination, it can efficiently detect the correct URIs for a given set of named entities within an input text. Furthermore, AGDISTIS is agnostic of the underlying knowledge base.
AGDISTIS has been evaluated on different datasets against state-of-the-art named entity disambiguation frameworks.
Moreover, FOX is a modern Named Entity Recognition Framework which uses AGDISTIS for the Named Entity Linking Part. Additionally, FOX is used by DEER to extract named entities from resource descriptions.
- Short Description of AGDISTIS - Framework can be found here.
- Supplementary material can be found here
- More evaluation of AGDISTIS can be found here.
For more information please visit our project site at Github.
09.02.2015 - Since our last measurement in September 2014 AGDISTIS language versions were used as follows:
- 88306 English AGDISTIS (before 63664 calls)
- 473 German AGDISTIS (before 301 calls)
- 258 Chinese AGDISTIS (before 168 calls)
This steady growth is amazing! Thank you all!
23.10.2014 - We won the Best Research Paper Award at ISWC 2014. Read more here: http://blog.aksw.org/2014/aksw-successful-at-iswc2014/
Here is the video from the closing ceremony.
21.10.2014 - We present the AGDISTIS demo at Stand 79 at ISWC 2014. Visit us! Discuss with us!
22.09.2014 - Try out our new demo of AGDISTIS: http://agdistis.aksw.org/demo * Updated to DBpedia 2014 * Faster CSS and JS
16.09.2014 - Short news: the English version of AGDISTIS has been called 63664 times since deployment in October 2013. Sofar the German and the Chinese endpoint of AGDISTIS have been called 301 respectively 168 times since July 2014.
15.09.2014 - New version released!
AGDISTIS is now more efficient, faster and especially easier to configure.
12.09.2014 - We are currently improving maintainability of AGDISTIS by doing a light version of it. For all experiments as described in "AGDISTIS - Graph-Based Disambiguation of Named Entities using Linked Data by Ricardo Usbeck, Axel-Cyrille Ngonga Ngomo, Sören Auer, Daniel Gerber und Andreas Both in International Semantic Web Conference" have a look at commit #80 or release v0.0.1
02.07.2014 - Since July 2014 we also provide a Chinese endpoint:
curl --data-urlencode "text='The <entity>shanghai</entity> in <entity>北京市</entity>.'" -d type='agdistis' http://titan.informatik.uni-leipzig.de:8080/AGDISTIS_ZH
and a German endpoint:
curl --data-urlencode "text='Die Stadt <entity>Dresden</entity> liegt in <entity>Sachsen</entity>.'" -d type='agdistis' http://titan.informatik.uni-leipzig.de:8080/AGDISTIS_DE