Click to edit Master title styleClick to edit Master text stylesSecond levelThird levelFourth levelFifth levelAn Efficient Approach to Clustering in Large Multimedia Databases with NoiseAlexander Hinn

clearN=200取200个数w(1)=0w=randn(1N)产生一个1×N的行向量第一个数为0w为过程噪声(其和后边的v在卡尔曼理论里均为高斯白噪声)x(1)=0状态x初始值a=1a为状态转移阵此程序简单起见取1for k=2:N x(k)=ax(k-1)w(k-1) 系统状态方程k时刻的状态等于k-1时刻状态乘以状态转移阵加噪声(此处忽略了系统的控制量)endV=randn(