Active Online Learning for Social Media Analysis to Support Crisis Management

Description

People use social media (SM) to describe and discuss various situations involving them, such as crises. It is therefore worth exploiting SM content to support crisis management, particularly by revealing in real time useful and unknown information about crisis. We are therefore proposing an active online classifier of multiple prototypes, called AOMPC. It identifies appropriate crisis information. AOMPC is an online learning algorithm operating on information streams and fitted with active learning systems to continually query the ambiguous unlabelled information label. A fixed budget strategy controls the number of queries. AOMPC accommodates partly labelled data streams. AOMPC was evaluated using two types of data: synthetic data and SM data from Twitter related to two crises, Colorado Floods and Australia Bushfires. To provide at thorough evaluation, a whole set of known metrics was used to study the quality of the results. A comparative study of AOMPC against other available online learning algorithms was performed. The experiments showed very good behavior of AOMPC for dealing with evolving, partly-labelled data streams.