Based on FCM Disaster Area Management at Different Levels

Abstract

This chapter mainly researches the classification result of the stricken areas based on the Fuzzy C-mean algorithm. The mortality, building collapse rate, and seismic intensity are considered as the reference factors which need Fuzzy C-Means clustering analysis firstly, and then the reference factors will be converted by nondimensionalizing process in order to compress all the data in the standard interval of [0, 1]. The thresholds of operation will be defined in the process of data initialization, if the condition of thresholds is realized, the operation will be stopped; secondly, iterations of algorithm will be stated. The consequence obtained after the stop of the initialization will be used to calculate the subjection degree of each factor again, and is divided into several intervals in order to construct a matrix set, and then cluster center will be calculated again. After the many steps of repetitive computations, if the cluster center and the matrix set meet the conditions of thresholds and the iterations which are both set by us, and the condition that thresholds is realized means the astringency of algorithm is outstanding enough to meet requirements of relevant target. Besides, the algorithm will be simulation analyzed using MATLAB software, and at last take the some regions such as Qingchuan County, An County and Mianzhu City for example to get the final result. Stricken areas are classified into three categories, and the result will be analyzed in the view of different proportions of all the factors.