Identifying Argumentative Discourse Structures in Persuasive Essays

Text mining project at Universitat Passau

tools

• 1.Python 2.7
• Python libraries
• Pandas
• Numpy
• Scipy
• gensim
• Scikit learn
• nltk library
• Pycharm or suitable IDE tool
• Brat reader ,Integrated with the project

Datasets Train and Test

• Christian Stab and Iryna Gurevych. 2014. Annotating argument components and relations in persuasive essays. In Proceedings of the 25th International Conference on Computational Linguistics (COLING 2014), Dublin, Ireland

• Christian Stab and Iryna Gurevych. 2016. Parsing Argumentation Structure in Persuasive Essays. arXiv preprint, https://arxiv.org/abs/1604.07370

Note

• Certain blocks of code has been commented since it needs to be compiled once , if Datasets are changed picklefiles needs to be trained again

Steps to Load and train your dataset

• Collect corpus which should be annotated as shown below in the link

• Example for annotated data

• Bratreader

• Using the above bratreader to extract each annotated essay as annotated object which contains several attributes which are used in Data preprocessing

  • Place the annotated files ".ann" at "ArgumentMining-Backend/bratessays/"

Usage

Usage Instructions

Metrics

Metrics sample

This project is only for education purpose only !!