@phdthesis{oai:tsukuba.repo.nii.ac.jp:00008701,
 author = {Orihara, Ryohei and 折原, 良平},
 month = {},
 note = {Desire to mechanize non-deductive reasoning has resulted some fruits in the field of artificial intelligence. On induction, classification rule learning systems have been widely studied. On analogical reasoning, a problem solver that uses a similar experience in the past was proposed. On abduction, an efficient mechanism to manage consistency of sets of hypotheses is developed. However the past sudies also have problems. The first issue is on knowledge description. A way to describe a task of non-deductive reasoning system can hugely influence effectiveness of the system. The second issue is on semantics. A uique application of abduction on logic program is seldom discussed, though its semantics is widely studied. In analogical reasoning, a desire to implement practial systems has somehow postponed analyzing semantic nature of analogical reasoning. The thesis attempts to propose solutions for the problems. For the first issue, preprocess system or a subsystem that can change the description of the task is given for the main reasoning system. For the second issue, analogical reasoning is regarded as an application of abduction and its declarative semantics is given. The main outcomes of the thesis are as follows.  ･･･, 1999, Bibliography: p. 97-102},
 school = {筑波大学, University of Tsukuba},
 title = {Knowledge description and semantics in non-deductive reasoning},
 year = {1999}
}