Algorithm 4: Fuzzy C-Means clustering

Input:x1…xl—objects, K—number of clusters , m—the degree of cluster fuzziness

Output: ${\left\{w_{ki}=w_k\left(x_i\right)\right\}}_{i=1,k=1}^{l,K}$—coefficient sets with the degree of xi belonging to the k-th cluster

Each object x is randomly assigned a cluster membership coefficient wk(x), k =1 ..K

while $\parallel w_{\left(t+1\right)}-w_{\left(t\right)}{\parallel }^2\ge \varepsilon \left(or\parallel C_{\left(t+1\right)}-C_{\left(t\right)}{\parallel }^2\ge \varepsilon \right)$ do

$c_k=\left[{\sum }_x{w}_k^m\left(x\right)x\right]/\left[{\sum }_x{w}_k^m\left(x\right)\right]$

$w_k\left(x\right)={\left[{\sum }_{c\in C}{\left(\parallel x-c_k\parallel x-c\parallel \right)}^{\frac{2}{m-1}}\right]}^{-1}$

return{wki}