A Web-Based Disaster Monitoring System On Tweets Using Methods Of Classification And Named Entity Recognition

Taqiuddin, Muhammad Ghifari (2025) A Web-Based Disaster Monitoring System On Tweets Using Methods Of Classification And Named Entity Recognition. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5025211063-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

This project develops a web-based system for monitoring disaster-related tweets in Indonesia using tweet classification and named entity recognition (NER). The system collects tweets based on selected disaster-related keywords, then classifies each tweet to assess whether it indicates an actual disaster event. A dataset of 178,648 labeled tweets, annotated as TRUE for real disaster events and FALSE otherwise, was used to train four classification models, with XGBoost achieving the best F1-score of 0.97114. For the NER task, four models were fine-tuned on an adjusted dataset derived from Ariyanto et al. (2025), containing annotated disaster entities. IndoBERT obtained the highest F1-score of 0.93491. These models are combined into a system that continuously retrieves recent tweets, identifies relevant ones, extracts key disaster-related information, and displays the results on a web interface. While effective overall, the system still struggles with ambiguous tweets, informal language, abbreviations, stylized writing, and recognizing disaster events outside Indonesia.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Named Entity Recognition, Text Clasification, Tweet Scraping, Web-based System
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Ghifari Taqiuddin
Date Deposited:	31 Jul 2025 03:02
Last Modified:	31 Jul 2025 03:02
URI:	http://repository.its.ac.id/id/eprint/124430

Actions (login required)

View Item