RBF神经网络拟合函数的，附注释

close all;
clc;

t1=[0:0.1:pi]‘;%自变量t1
y=sin（t1）;%因变量y
size_n=size（t11）;%输入样本的个数
train_num=10;%训练样本个数

hid_num=5;%隐含层神经单元个数
w=0.3*zeros（1hid_num）;%隐含层-输出层的权值
b=0.8*ones（1hid_num）;%节点基宽
% center=[2.490.831.693.25-0.02]; %基函数中心 sin pi
% center=[6.862.4018.241.644.63]; %基函数中心 sin 2pi
center=rand（1hid_num）;
eta_w=0.25*ones（1hid_num）;%学习率η
eta_b=0.25*ones（1hid_num）;
eta_c=0.25*ones（1hid_num）;
alpha=0.5;%动量因子α

maxgen=800;%最大迭代次数
t2=1:1:maxgen;
error_goal=0.000000000000000000000000000000000000000001;%误差精度要求
mse=zeros（1maxgen）;
wk_1=w;wk_2=w;%前1 2 时刻的权值
bk_1=b;bk_2=b;%前1 2 时刻的基宽值
ck_1=center;ck_2=center;%前1 2时刻的中心值

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  提取训练样本  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
for i=1:train_num
    seq=floor（i/train_num*size_n）;
    train_in（i）=t1（seq:）;
    train_out（i）=y（seq:）;
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  迭代学习参数  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
gen=0;  
for i=1:maxgen
    gen=gen+1;
    B（gen:）=b;
    W（gen:）=w;
    C（gen:）=center;
    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%  训练样本  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
    for j=1:train_num
        for k=1:hid_num
            h（k）=exp（-norm（train_in（j）-center（k））^2/（b（k）^2））;
        end
        ye（j）=w*h‘;
        %%%%%%%%%%%%%%%%%%%%%%%%  隐含层参数学习  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
        for k=1:hid_num
            e=train_out（j）-ye（j）;  %定义误差     
            w（k）=wk_1（k）+eta_w（k）*e*h（k）+alpha*（wk_1（k）-wk_2（k））; %隐含层-输出层的权值w学习公式   
            b（k）=bk_1（k）-eta_b（k）*（e*w（k）*h（k