DKPro Argumentation

Java framework for demonstration purposes of working with DKPro/UIMA typesystem for argumentation annotation and argumentation mining.

Requirements

Java 1.8 and higher

Installation

DKPro Argumentation is available from Maven Central. For working with the UIMA type system, add the following dependencies to your pom.xml

<dependency>
    <groupId>org.dkpro.argumentation</groupId>
    <artifactId>dkpro-argumentation-types</artifactId>
    <version>0.1.0</version>
</dependency>

• Current development version: 0.2.0-SNAPSHOT

Modules

org.dkpro.argumentation.types

The UIMA type system is the central part of DKPro Argumentation. It allows one to annotate argument components and relations on top of the UIMA/DKPro framework.

The class hierarchy contains two central classes, ArgumentComponent and ArgumentRelation.

IO

Contains ArgumentDumpWriter class that debugs all argument annotations in a document to the file/stdout.

Misc

Provides classes to easily operate with argument component annotations.

Preprocessing

Contains several annotators for casting the problem of argument component identification as BIO tagging (i.e. ArgumentTokenBIOAnnotator)

Examples

Example project for reading annotated data

as shown in

Habernal, I., Eckle-Kohler, J., & Gurevych, I. (2014). Argumentation Mining on the Web from Information Seeking Perspective. In E. Cabrio, S. Villata, & A. Wyner (Eds.), Proceedings of the Workshop on Frontiers and Connections between Argumentation Theory and Natural Language Processing (pp. 26-39). Bertinoro, Italy: CEUR-WS. Retrieved from http://ceur-ws.org/Vol-1341/

Requirements

• Annotated data
  • packaged separately, available at https://www.ukp.tu-darmstadt.de/data/argumentation-mining/argument-annotated-user-generated-web-discourse/

How-to

1. Modify paths to gold data

• Modify org.dkpro.argumentation.tutorial.ArgumentationCorpusDebugger and set the annotatedCorpusDir variable to point to the gold data located in gold.data.toulmin directory

2. Run ArgumentationCorpusDebugger

• It will print annotated argument components, relations, and other info to the std. out

3. Explore it further!

• Have a look at org.dkpro.argumentation.io.writer.ArgumentDumpWriter from the org.dkpro.argumentation.io.writer package which shows how to access the argument components, their text, tokens, sentences, etc.

© 2013-2017 UKP

Argumentation Mining Special Interest Group members

https://www.ukp.tu-darmstadt.de/ukp-home/research-areas/argumentation-mining/