search

joelansbro / nlp-methods

2 stars 0 forks source link

NLP: Argumentation Mining Investigation #5

Open joelansbro opened 2 years ago

joelansbro commented 2 years ago

Argumentation Mining is the NLP method of exploring and mapping the arguments presented in a piece of text. This method could be useful in generating a map of an argument that an article makes, providing some useful insight when researching the blogs.

I should use the existing datasets inside the NLP-methods repo to test out the argumentation mining, look up previous examples of argumentation mining and see if I can replicate something that looks half decent.

Some things to read:

Five Years of Argument Mining: A Data-Driven Analysis 2018 Elena Cabrio, Serena Villata Argument Mining: A Survey 2020 Lawrence, Reed Argumentation Mining State of the Art and Emerging Trends 2016 Marco Lippi, Paolo Torroni

Give this a go: https://towardsdatascience.com/cross-topic-argument-mining-learning-how-to-classify-texts-1d9e5c00c4cc

joelansbro commented 2 years ago

I've added a python package that parses some arguments from given sentences, from an academic source. It's a good basis to begin off of. See notes section for further details.

Main thing - can we gives weight to the argument categories to define what a good argument is? If so, can we add this into the weighting of what defines a good article?

joelansbro commented 2 years ago

Created a prototype testing out Argument Mining, the next step will be to run a few of the articles over the argument mining in order to get a feel for how good the argument mining has gone, Get a feel for how well the argument mining scores have captured a blog's points, and present the numbers. ending with a summary of findings (can be rough notes) that will be able to lead into explaining in the project.