一种新城市气温模式分类的聚类算法

城市气温是对城市气候特性评价的一个重要指标.提出核概率聚类算法并将其应用于城市气温的模式分类中,以此寻找城市发展上的共同点.该算法在概率聚类算法上引入了核学习方法的思想,能够很好地处理噪音和孤立点,实现更为准确的聚类.实验结果表明,与相关聚类算法相比,核概率聚类算法聚类效果好,且算法能够很快地收敛.

Kernel-based Possibilistic Clustering Algorithm Applied to Pattern Classification of City Temperature

City temperature is an important index to evaluation of city climate feature.Kernel-based possibilistic clustering algorithm proposed in this paper is applied to pattern classification of city temperature so as to find common ground on city development.The algorithm combines the idea of kernel learning method with possibilistic clustering algorithm.It can preferably deal with noise and outliers and accurately realize clustering. Compared to other relevant algorithms,the simulation results show that the algorithm has the advantage of good clustering effect and fast convergence.