Question

matlab中fcm算法可以对一维数据进行分类吗

不知道这个data在matlab中怎样输入





Answer 1

function [U,V,iteration] = std_fcm(X,c)
% std_fcm:standard fcm by liyang @BNU Math 315
% Email:farutoliyang@gmail.com
% 2009.09.25
% input:
% [num_sample,num_attribute] = size(X) let N = num_sample
% X = (x(1);x(2);...;x(num_sample));
% c:classnumber
% output:
% U : c*num_sample
% V = (v(1);v(2);...;v(c)) : c*num_attribute
% Problem:
% min Q(U,V) = sum(i=1,...,c)sum(k=1,...,N)( u(i,k)^2 * distance(x(k),v(i))^2 )
% subject to sum(j=1,...,c)u(j,t) = 1, for each t = 1,2,...,N
% solve in the Euclidean distance sense
% u(s,t) = 1 / ( sum(j=1,...,c)(distance(v(s),x(t))/distance(v(j)-x(t)))^2 )
% v(s) = ( sum(k=1,...,N)( u(s,k)^2*x(k) ) ) / ( sum(k=1,...,N)( u(s,k)^2 ) )
[num_sample,num_attribute] = size(X);
N = num_sample;
%% initialization
epsilon = 0.001;
iteration = 1;
U = rand(c,N);
V = zeros(c,num_attribute);
%% 主体循环
while(1)

% calculate new V
for s = 1:c
temp_numerator = 0;
for k = 1:N
temp_numerator = temp_numerator + U(s,k)^2 * X(k,:);
end
V(s,:) = temp_numerator ./ sum( U(s,:).^2 );
end

% calculat new U
for s = 1:c
for t = 1:N
temp_denominator = 0;
for j = 1:c
temp_denominator = temp_denominator + ( ED(V(s,:),X(t,:))/ED(V(j,:),X(t,:)) )^2;
end
new_U(s,t) = (temp_denominator)^(-1);
end
end

% 主体循环终止条件
if max(max(abs(U-new_U))) < epsilon
break;
end

U = new_U;
iteration = iteration + 1;
end

%% Euclidean distance function
function d = ED(x,y)
d = sum((x-y).^2).^0.5;

追问

这个，，，我是菜鸟级别，没看懂啊
我只是想问我这个data这个参数应该怎么输入啊