核密度估计

具体原理及定义：传送门 https://en.wikipedia.org/wiki/Density_estimation

MATLAB 代码实现如下：

 1 % Kernel Density Estimation
 2 % 只能处理正半轴密度
 3 function [t, y_true, tt, y_KDE] = KernelDensityEstimation(x)
 4 % clear
 5 
 6 % x = px_last;
 7 % x = px_last_tu;
 8 %%
 9 %参数初始化
10 Max = round(max(x));           %数据中最大值
11 Min = round(min(x));           %数据中最小值
12 Ntotal = length(x);     %数据个数
13 tt = 0 : 0.1 : Max;     %精确x轴
14 t = 0 : Max;            %粗略x轴
15 
16 y_KDE = zeros(10 * Max+1, 1);   %核密度估计值
17 sum1 = 0;                       %求和的中间变量
18 %%
19 %计算带宽h
20 R = 1/(2*sqrt(pi));
21 m2 = 1;
22 h = 3;
23 % h = (R)^(1/5) / (m2^(2/5) * R^(1/5) * Ntotal^(1/5));
24 
25 %%
26 %计算核密度估计
27 for i = 0 : 0.1 : Max        
28     for j = 1 : Ntotal
29         sum1 = sum1 + normpdf(i-x(j));
30     end
31     y_KDE(round(i*10