This repository contains the code and data for an approach to the CheckThat! 2024 Task 2: Subjectivity Classification. The task is to determine whether a sentence from a news article is subjective or objective. The system in this repository is specifically designed for the English language.
data/: Contains the training and test data.main.ipynb: The main notebook that contains the code for the system.evaluation/: Contains the score and metrics for different models and parameters.evaluation.tsvcontains the score for different models, preprocessing techniques, and features.paratuningLinReg.tsvcontains the scores for parameter optimization of the Logistic Regression model.paratuningSVM.tsvcontains the scores for parameter optimization of the SVM model.resultTest.txtcontains the scores and metrics for the dev_test and test data. Both datasets have only been used for the final evaluation of the system and were only run once.documentation/: Contains slides documenting the system.
To reproduce the results stated in the documentation you just have to run main.ipynb.