Research And Application Of Time Sequence Comparability Matching Algorithm In Decision Support System Of Storm Surge

China has vast ocean area and contains enormous developable resources, which canbe of great importance in China’s economic development, as well as the global business.However, the rapid development of global economy and the frequent climate change hasresulted in global ocean acidification, raising the sea level and marine disasters. And the placeour country is located in is a special geographical location which makes our country highlyvulnerable to extreme weather and the impact of ocean processes, such as storm surges, bigwaves and so on. The storm surge leads to unusual increasing in sea levels and causes the sea tosurge inland, and flooding disaster occurs. Therefore, using the scientific computer technologyfor data mining of the natural disasters based on valuable historical information and developingreasonable and effective solutions has important meaning to the construction of disasterprevention and mitigation, disaster evaluation，and the decision support of the governmentagencies.The paper applies the concepts and algorithms of time series similarity matching tothe disaster classification in decision support system of storm surge, and combining withdomain knowledge, presents a new time series similarity matching algorithm for storm surge.The main study contents are as follows:(1) Distance measure model for time series similarity matching. Distance measure isthe key point to the matching algorithm, The distance measure model for the storm surge timeseries similarity matching is presented the time series in storm surge, storm surge, decisionsupport, based on the definition of storm surge time series, decision support of storm surge andthe storm surge time series similarity;(2) The selection of dimension-reduction method. According to the characteristics of the storm surge data and the analysis of dimensionality-reduction methods proposed bydomestic and foreign experts in their respective fields, a dimension-reduction method isproposed based on the slope of segments;(3) A similarity matching method is proposed. First, we define the time series ofstorm surge, similarity measure model of storm surge, then the storm surge time series arefiltered with the segmentation algorithm based on slope feature vectors and finally, the similarsequences are extracted by similarity measure model of storm surge. According to theinformation of the extracted sequences in the database, decision support of storm surges can becarried out;(4) The implementation of the decision support system of storm surge. First, thesimilarity matching method based on the slope of the feature vector is used in disasterclassification sub module, which is integrated with the sub modules of the decision supportsystem using the middleware Enterprise Service Bus (ESB) technology and the storm surge hasbeen integrated support system, which achieves the system requirements of portability andreusability well. Finally the technology of Flex and Java is used for the system, and the systemis user-friendly and easy for decision-makers to make intuitive decisions.The key points of this paper are the presentation of similarity matching methodapplied to storm surge disaster data based on the slope of the feature vector, the construction ofdisaster classification sub module and the development of the entire storm surge based on theEnterprise Service Bus (ESB).The combination of time series similarity matching and the decision support systemcan make full use of historical data and promote the accuracy, precision of disasterclassification, which provides good technical support for the study of storm surge disaster andall the ocean data.