A New Disaster Recognition Algorithm Based on SVM for ERESS: Buffering and Bagging-SVM

摘要

We have previously proposed an Emergency Rescue Evacuation Support System (ERESS). ERESS is based on Mobile Ad-hoc network (MANET) and aims to reduce the number of victims in panic-type disasters. ERESS consists of mobile terminals with various sensors. The former ERESS recognizes a disaster outbreak by basic SVM of machine learning. But its recognition accuracy is not so high and it needs a lot of computational effort. In this paper, we propose a new algorithm named as Buffering and Bagging SVM (BB-SVM). In this method, an ERESS mobile terminal accumulates the information about disaster outbreak by its buffer. Then, the terminal selects appropriate data from the buffer for disaster recognition. We show that the proposed method recognizes disaster outbreak accurately and quickly by panic-type experiments.