Effectively Handling Crisis Management through Social Media

Abstract

Individuals utilize social media (SM) to depict and talk about various circumstances they are engaged with, like emergencies. It is along these lines advantageous to abuse SM contents to help crisis the executives, specifically by uncovering valuable and obscure data about the emergencies progressively. Subsequently, we propose a novel active online multiple-prototype classifier, called AOMPC. It recognizes pertinent data identified with a crisis. AOMPC is an online learning algorithm that works on data streams and which is outfitted with active learning mechanisms to actively inquiry the mark of equivocal unlabeled data. The quantity of questions is constrained by a fixed spending procedure. Ordinarily, AOMPC obliges somewhat marked data streams. AOMPC was assessed utilizing two sorts of data: (1) synthetic data and (2) SM data from Twitter identified with two emergencies, Colorado Floods and Australia Bushfires. To give an intensive assessment, an entire arrangement of realized measurements was utilized to contemplate the nature of the outcomes. Additionally, a sensitivity examination was led to show the impact of AOMPC's parameters on the precision of the outcomes. A similar investigation of AOMPC against other accessible online learning algorithms was performed.