Case grammar which explains the grammaticality of a language and its semantics is used vastly in linguistics and natural language processing in computer science. However, the theory of case classification which is the basis of case grammar is not established generally.
In this thesis research, a case classification was proposed according to the frequency of the particular case element in the sentences of computer science literatures. And it corresponds to the types of questions( by W. LEHNERT ) with respect to the points required in the questions. The number of cases and computational efficiency is tradeoff relation, and for computational efficiency some expressiveness is sacrificed by the criterion of relative frequency.