Hierarchical and fuzzy C-means clustering

I computed hierarchical clustering and the best classification requires k=8 clusters. I wanted to find the probability of belonging to a cluster for each unit so I used fuzzy C-means clustering with k=8 but the probabilities that come out are different from the results obtained with hierarchical clustering.

Eg. According hierarchical cl. the first unit belongs to the first cluster while according to fuzzy belongs to:

• first cluster: probability = 0.25
• second cluster: pr = 0.46
• other clusters: pr 0.1

why this discrepancy? the input matrix is the same.

Can I somehow get the probability of belonging to a cluster? Any suggestions?

Thanks!