A novel spatio-temporal similarity algorithm adapted to typhoon disaster cases

Costal areas in China frequently suffer typhoon disasters, which always causes lots of damage. Learning from appropriate historical cases can improve the efficiency of emergency decision making. A novel spatio-temporal similarity algorithm is proposed in this paper in order to find the most relevant typhoon cases. Cases similar to current typhoon disaster can be retrieved and ranked by similarity. Many factors including meteorological attributes of a typhoon and complex information of affected areas are analyzed. The track, the wind velocity, and the year of typhoons are selected as the main factors for algorithm designing. At last, the feasibility of this algorithm is testified with historical data in recent years. The proposed method can also be applied to other disaster domains.