Big Data Challenge - Indonesian Election Twitter Classification in 2024

Overview

This project was developed by Nicholas Andrew's team for the Big Data Challenge, leveraging Natural Language Processing (NLP) techniques to classify Twitter comments related to the 2024 Indonesian election into 8 categories. The dataset was provided by Satria Data and contains a diverse range of opinions and discussions regarding the election. To achieve accurate classification, we implemented TensorFlow and PyTorch frameworks and experimented with multiple models, including Logistic Regression, BERT (Bidirectional Encoder Representations from Transformers), and LSTM (Long Short-Term Memory networks).

Key Techniques

1. Multi-Model Approach: Utilized a combination of Logistic Regression, BERT, and LSTM for improved classification performance.
2. Deep Learning Frameworks: Implemented TensorFlow and PyTorch to enhance NLP-based text processing and classification.
3. Optimized Model Performance: Experimented with hyperparameter tuning and different architectures to improve classification accuracy.