Development of a methods for analysis of uncertain spatial data on critical events

为了识别关键事件的空间趋势，提出了一种以关联点分布为中心的空间数据分析方法。利用核密度估计方法，根据已知点和未知点的分布构造了两个概率密度函数，并将两个概率密度函数之间的差表示为积分平方误差。通过最小化已知点和未知点分布之间的空间差异，提出了一种用未知点近似已知点分布的方法，并给出了利用点数据的求解方法。

A method for spatial data analysis focused on distributions of incidence points is developed to recognize the spatial trend in critical events. Two probability density functions are based on the distributions of the known points of the past and unknown points by kernel density estimation method, and the difference between these functions is formulated as integrated squared error. A method that approximates the distribution of known points by unknown ones and its solution using point data are proposed by minimizing the spatial difference between the two distributions.