Natural Disaster Monitoring Information System from Social Media Data Using Naïve Bayes Algorithm

In Indonesia, there have been several natural disasters, such as earthquakes, tsunamis, landslides, floods, and others. Because Indonesia is situated where the Eurasian, Pacific, and Indo-Australian plates converge, this potential natural disaster is caused by this location. Social media information is expanding quickly and becoming more useful. Social media helps to alert people of the disaster’s location during a disaster like a flood. Twitter is used as a data search engine in this work. Twitter has been utilized effectively to update the public on current events during emergencies. In order to learn more, we can conduct a search using pertinent hashtags to determining for the incident’s location. The test’s results will show a map of the Indonesian region, and the disaster’s epicenter will be determined using the geolocation provided by the tweet data. The Naive Bayes approach will be used for classification. The clustering process occurs in real time across every region of Indonesia. In this investigation, the accuracy value was 75% based on the k-fold cross-validation test, utilizing a fold value of 3.