Argument Quality Assessment

This repository contains Python code for automatically assessing the quality of argumentative documents, specifically focusing on student essays on the confirmation bias dimension. The code implements a feature-based supervised approach using text features and machine learning techniques.

Learning Goals

• Assessing the quality of argumentative documents automatically using text features and ML techniques.

Tasks

Develop a feature-based supervised approach to automatically assess the quality of arguments in student essays on the confirmation bias dimension.

• Confirmation Bias: The absence of opposing arguments.

Usage

1. Clone this repository:

   git clone <repository-url>

2. Run the Python script:

   python main.py

3. After execution, the predictions will be saved in a file named predictions.json.

Requirements

• Python 3.x
• NLTK (nltk)
• Pandas (pandas)
• NumPy (numpy)
• Scikit-learn (scikit-learn)

Note

• Adjust paths and parameters as necessary.
• Ensure that the input data files (essay-corpus.json and train-test-split.csv) are placed in the data directory.
• This script utilizes Support Vector Machine (SVM) for classification tasks.