Back to Showcase
The Bot Comment Detection System Using IndoBERT is a Natural Language Processing (NLP)-based machine learning project developed to automatically classify comments as either bot-generated or human-written. The system utilizes IndoBERT, a pre-trained Bidirectional Encoder Representations from Transformers (BERT) model specifically designed for the Indonesian language, to achieve high classification performance.
The project covers the complete machine learning pipeline, beginning with web scraping, followed by text preprocessing, dataset preparation, model fine-tuning, and performance evaluation. A balanced dataset consisting of 1,000 Indonesian-language comments was collected from six major online platforms: Kompas, Kompasiana, Detik, CNN Indonesia, Quora, and Kaskus.
Project Summary
The project covers the complete machine learning pipeline, beginning with web scraping, followed by text preprocessing, dataset preparation, model fine-tuning, and performance evaluation. A balanced dataset consisting of 1,000 Indonesian-language comments was collected from six major online platforms: Kompas, Kompasiana, Detik, CNN Indonesia, Quora, and Kaskus.
System Galleries