Algorithm 1: Silhouette Analysis (SA)

For every sample   1.Find the mean distance from all features in the same cluster $\left({𝒶}_i\right)$   2.Find the mean distance from all features in the closest cluster $\left({𝒷}_i\right)$   3. Find the coefficient: ${𝒮}_{𝒸}=\frac{{𝒷}^i-{𝒶}^i}{max({𝒶}^i,{𝒷}^i)}$ If${𝒮}_{𝒸}=0$, the sample is very close to the neighboring clusters If${𝒮}_{𝒸}=1$, the sample is far away from the neighboring clusters If${𝒮}_{𝒸}=-1$, the sample is assigned to the wrong clusters