Q-DAR: quick disaster aid and response model using Naïve Bayes and Bag-of-Words algorithm

Abstract

A real-time approach in decision-making for disaster management here in the Philippines is important, given that the country is vulnerable to disasters. Many studies show that data from social media can be of good use especially in times of disaster. This research aims to develop a model that will serve as a decision support tool for the government to respond during disasters with the use of Twitter hashtags that are based from the DRRM disaster phases; Disaster Response, Disaster Preparedness, Disaster Rehabilitation and Recovery, and Disaster Mitigation. The result of the study provides an overview of the critical level of the disaster phases with the use of Naïve Bayes and Bag-of-Words algorithm. The process of this study has four phases; the Training phase includes the labeling of the hashtags and list of Bag-of-Words, the next phase is Data Collection where Twitter data automatically updates and can be extracted every hour. The third phase is Data Pre-processing; where the tweets go through tokenization and normalization. Lastly, the dataset goes through Q-DAR with two sub-processes: The Disaster Phase Analysis using Bag-of-Words and Critical Level Analysis using Naïve Bayes in WEKA.