Methods in fuzzy logic have been applied to serve as secondary classifier f
or a hierarchical classification model. The use of this model in interpreta
tion of mammograms is discussed. Also is discussed, the inevitability of us
ing a fuzzy approach in the problem. Finally, the two different fuzzy appro
aches for secondary classification are compared on basis of their performan
ce as far as clustering is concerned. The idea of using a fuzzy covariance
matrix [5,6] in the distance metric of the classical c-means algorithm [1-3
] has also been tried. (C) 2000 Published by Elsevier Science B.V. All righ
ts reserved.