Machine Learning Based Automated Disaster Message Classification System Using Linear SVC Algorithm

Abstract

This paper presents the machine learning based automated disaster message classification system. Machine learning is be used to identify such information and provide valuable information for aiding disaster response during emergency events. Disaster management prediction systems (DMPS) are computer systems for determining when and where to deploy mitigation measures in the event of an emerging natural or man-made hazard, while accounting for and mitigating human factors that may compromise operational effectiveness for providing fast services to handle this high volume and velocity of urgent information. Till date diverse techniques used for disaster and pandemic management are available using the technologies like satellite-based systems, cellular networks, Internet of things (IoT), smartphone-based systems, 5G and cellular networks. Linear support vector machines are an efficient way to learn discriminative models, which is especially useful in data where the number of attributes is large or is not known. Linear SVC (Support Vector Classifier) is one of the most successful linear models, not only because it is quite fast to train and compute, but also because it can achieve excellent performance in high dimensional problems. Linear SVMs can make use of a wide variety of learning algorithms.The proposed work uses Linear SVC Algorithm with strategy of self-training that learns from available datasets with the labeled data. Finally, the paper gives the message classification based on the emergency to the relevant disaster.