The Research On Comprehensive Assessment And Loss Prediction Methods Of The Disaster Typhoon And The Development Of Its Software Platform

China is the one of the countries that who suffers serious influence of the typhoon disaster. The typhoon has brought huge economic losses to our country, endangered the safety of people’s lives, and did some bad influence to the social stability. Therefore, the research of estimating the integrated risk index and predicting the loss of typhoon take high station in reducing the harm of the typhoon.Zhejiang province is located in the southeast coast of china. He has long coastline and wide oceans. So he suffered much from the typhoon disaster. In this paper, we do some research on the typhoon disaster from qualitative and quantitative angles according to the records provided by Zhejiang provincial meteorological bureau. The qualitative analysis can be divided into disaster assessment and disaster pre-assessment. The steps of evaluating the instance of typhoon disaster damage are grade, data process and assessment. Based on the concept of integrated risk index, we use the loss index as the reference sequence and use the theory of grey relevancy method to evaluate the integrated risk index after a comprehensive consideration of the disaster causing factors and the disaster bearing bodies. The index reflects how much the typhoon brings to the disaster bearing bodies. Then we use the same theory to evaluate the instance of typhoon disaster damage. It is proved that the integrated risk index and the disaster damage index are corroborate-ons to each other.With the development of information science, the qualitative evaluation can’t meet the growing needs of peoples. So we use the method of GRNN and SVM to do the typhoon disaster prediction before the disaster coming. The purpose is to testing their feasibility of predicting. The result shows they are feasible to predict the loss of typhoon disaster. The fitting of historical data all achieves90%. And the forecast accuracy also reached80%. Take a single prediction model, SVR is more reliability, but it runs slower. GRNN has strong nonlinear mapping ability, it can remember all the samples, but makes more errors while predicting.Finally, according to the needs of the business, we constructed a research platform. Its name is the platform of integrated risk estimating and loss predicting for the typhoon disaster. We use B/S structure, java language, SQL server database, and the powerful computational tool matlab. Then we can operate the complicated data by the integrated platform. The platform can assess and predict, can do qualitative evaluation and quantitative prediction for the typhoon disaster, can evaluate and predict before the typhoon comes and can assess after the typhoon has happened. The research can provide decision to the related departments of disaster prevention and mitigatio-n. It has important significance and can guide the disaster prevention and reduction of typhoon, then protects the national economic and to maintain the safety of people’s life and property.