Mantıksal ilişkilerin dizilerin özgül genellemesine dayanan bir yöntemle tümevarımsal çıkarılması

Learning logical relations from examples expressed as first order facts has beenstudied extensively by the Inductive Logic Programming research. Learning withpositive-only data may cause overgeneralization of examples leading to inconsis-tent resulting hypotheses. A learning heuristic inferring specific generalizationof strings based on unique match sequences is shown to be capable of learningpredicates with string arguments. This thesis outlines the effort showed to buildan inductive learner based on the idea of specific generalization of strings thatgeneralizes given clauses considering the background knowledge using least gen-eral generalization schema. The system is also extended to generalize predicateshaving numeric arguments and shown to be capable of learning concepts such asfamily relations, grammar learning and predicting mutagenecity using numericdata.Keywords: indective logic programming, machine learning, string generalization,hypotheses, example, background knowledge.iii