Conceptual clustering of structured objects: a goal-oriented approach

摘要: Abstract Conceptual clustering is concerned with problems of grouping observed entities into conceptually simple classes. Earlier work on this subject assumed that the and classes are described in terms a priori given multi-valued attributes. This research extends previous three major ways: • - characterized as compound objects requiring structural descriptions. relevant descriptive concepts (attributes relations) not necessarily but can be determined through reasoning about goals classification. inference rules used to derive useful high-level from initially provided low-level concepts. The created using Annotated Predicate Calculus (APC), which typed predicate calculus additional operators. Relevant appropriate for characterizing by tracing links Goal Dependency Network (GDN) represents relationships between goals, subgoals, related An experiment comparing results program cluster/s implements classification generation process obtained people indicates proposed method might offer plausible cognitive model processes well an engineering solution automatic generation.