en-es
en-fr
en-sl
en
0.25
0.5
0.75
1.25
1.5
1.75
2
Sustainable Linked Data generation: the case of DBpedia
Published on 2017-11-281051 Views
dbpedia ef, the generation framework behind one of the Linked Open Data cloud’s central interlinking hubs, has limitations with regard to quality, coverage and sustainability of the generated dataset.
Related categories
Presentation
Sustainable Linked Data Generation The case of DBpedia00:00
DBpedia00:13
DBpedia describes00:30
DBpedia contains00:38
DBpedia keeps growing, but there are quality issues00:46
There are 2 types of quality issues00:59
What are the causes?01:23
DBpedia’s Extraction Framework extracts Wikipedia’s infoboxes01:25
The community created mapping rules from infobox properties to a schema01:43
Causes for the issues can be found in the Extraction Framework (EF)01:53
Causes for the issues can be found in the mapping rules (MR)02:00
Causes for the issues can be found in Wikipedia itself02:02
Our goal is to adjust the EF & the MR to provide a more sustainable framework02:12
We integrated a generic, modular, and sustainable mapping language02:21
The result is a framework that enables sustainable Linked Data generation02:35
DBpedia is making the switch!02:44
Before02:59
Hard-coded mapping rules03:31
Hard-coded mapping rules - 103:41
Hard-coded mapping rules - 203:52
No machine-interpretable MR04:43
No machine-interpretable MR - 104:56
No machine-interpretable MR - 205:02
Restricted to the DBpedia ontology - 105:27
Restricted to the DBpedia ontology05:51
Limitations of the EF06:18
No schema validation06:23
An infobox template for defining persons on Wikipedia06:57
> 253 000 pages use the “person” infobox template07:11
Only one mapping is responsible for extraction07:18
Changes at least 250 000 times07:23
Wrong at least 250 000 times07:33
No schema validation - 107:52
No schema validation - 208:00
After08:10
Our solution08:38
Our solution - 108:56
Declarative mapping rules09:11
Declarative mapping rules -109:19
Declarative mapping rules - 209:21
Declarative mapping rules - 309:23
Machine-interpretable format09:36
Machine-interpretable format - 109:40
Schema validation10:09
A sustainable framework is needed that provides11:04
Usage of other ontologies11:17
Progress11:26
Switching to RML11:43
Evaluation: coverage11:58
Evaluation: performance12:16
Evaluation: flexibility12:36
Evaluation: flexibility - 112:50
A sustainable framework that has13:22
The future is bright!13:44
Made possible by14:10