- This repository contains code that has different entity linking tools and performance of those tools against wikipedia ground truth.
- Implementation of different algorithms for Entity Linking and ranking from the paper "Entity Linking with a Knowledge Base: Issues, Techniques, and Solutions by Wei Shen, Jianyong Wang, Jiawei Han"
- Java 1.8
- Maven
-
Download the zip or clone the entire repository.
-
Go to the project folder. [Note:-] You have to cd inside EntityLinkerTools
-
#type mvn package [Note:-] You have to cd inside EntityLinkerTools, where you can see pom.xml
-
Mavel will take care of everything, you should see Build success.
-
Now go to the target folder. You wil find a jar.
-
Run the jar by typing
-
java -jar [jarname] corpus (Note: without square[] bracket)
-
Here corpus is the directory for the LeadParagraphcorpus.
Always run the jar which has "with dependencies". In my case it is EntityLinkerTools-0.0.1-SNAPSHOT-jar-with-dependencies.jar
Here corpus is the leadParagraph.cbor
- You will see the F1 Measure for DBpedia Spotlight against TREC-CAR dataset in the command window
If you are facing issues with Maven build, I have placed the jar in the repository. You can directly execute the jar by typing java -jar EntityLinkerTools-0.0.1-SNAPSHOT-jar-with-dependencies.jar "path/to/corpus"
will modify Tagme tool soon. and more results set comming.