The implementation of the Naive Bayes Algorithm was utilized to detect tweets related to disasters

Abstract

Social media has become an important part of everyday life, with Twitter being a popular micro-blogging and social networking platform that allows users to share news, information, and personal thoughts. In times of emergencies or disasters, Twitter has proven to be a crucial communication medium. The widespread use of smartphones and tablets enables individuals to report emergencies in real-time, which can potentially save countless lives by alerting others to take necessary precautions. Several organizations are attempting to analyze tweets programmatically to detect disasters and emergencies, which can be beneficial to millions of internet users by providing timely alerts in the event of a crisis. However, the challenge lies in distinguishing between tweets related to a disaster and those that are unrelated. Twitter data is unstructured, making it necessary to use Natural Language Processing (NLP) to classify tweets as either "Related to Disaster" or "Not Related to Disaster". This research paper focuses on building a Naïve bayes classifier model and evaluating its accuracy by predicting on a test set created from the original dataset.