Clustering techniques have been used for data abstraction. Data abstra
ction has many applications in the context of databases. Conceptual mo
dels are used to bridge the gap between the user's view of a database
and the physical view of the database. Semantic models evolved to over
come the limitations of classical data models such as network and rela
tional models. The paper uses a knowledge-based clustering algorithm t
o extend the abstractions, such as classification and association, whi
ch are employed in the semantic modeling of databases. The complexity
of the proposed clustering algorithm is analysed. The ''tended semanti
c model can be used to design databases in which useful and interestin
g queries can be answered. The efficacy of the proposed knowledge-base
d clustering approach is examined in the context of a library database
.