Introduction

This repository contains the source code for the Joint Emotion and Cognitive Topic Model (JECTM) and BERT-based classification models as described in the associated manuscript.

Joint Emotion and Cognitive Topic Model (JECTM)

The JECTM model is implemented in Java.

Key Files:

• JECTM/Test/JECTM.java: Main entry point for running the JECTM model
• JECTM/Model/JECTM_Model.java: Core implementation of the JECTM model
• JECTM/eval/CoherenceEval.java: Code for evaluating topic coherence

Usage:

To run the JECTM model, execute the JECTM.java file. Make sure to set up the required data and parameters as specified in the code comments.

BERT-based Classification

This section implements BERT and other transformer-based models for emotion and cognitive engagement classification.

Key Files:

• training.py: Main script for training the classification models
• train_eval.py: Contains training and evaluation functions
• predict_pretrained.py: Script for inference using trained models

Training and Evaluation:

To train a model:

python training.py --model <Bert>

Inference:

To run inference on new data:

python predict_pretrained.py --model <Bert>

This script will load a trained model and make predictions on the data specified in the script.

Data

The models expect data in specific formats:

• JECTM: Excel files containing learner achievement and discussion data
• BERT: CSV or Excel files with text content and labels

Refer to the individual script comments for more details on data formatting and paths.

Requirements

• Java (for JECTM)
• Python 3.8
• PyTorch
• Transformers library
• Pandas, NumPy, and other common Python data science libraries

Citation

If you use this code in your research, please cite our paper:

[Include citation information here]

License

[Specify the license under which this code is released]