Classification and Prediction of Economic Losses - Storm Surge Disasters in Guangdong Province of China

Abstract

This paper took Guangdong province as an example, using the statistical data of twenty times storm surges from 2003 to 2010 to evaluate the disasters and predict the economic losses. We expected it to supply with sound references and proof for the decision-makers to prevent storm surges. With economic indices of direct economic losses, collapsed houses, damaged farmland area, et al., this paper used entropy method and factor analysis method to grade the storm surges into separate levels, which are the mild disaster, the moderate disaster, the serious disaster and the extra serious disaster. By BP neural networks and gray prediction method, we established the evaluation and prediction models of direct economic losses. Comparing the results of both methods, it found that neural network is more applicable and accurate to predict the economic losses of storm surges.