著者

近藤 誠一 山本 章博

出版者

人工知能学会

雑誌

人工知能学会全国大会論文集 (ISSN:13479881)

巻号頁・発行日

vol.27, 2013

---

We present a method to solve Inductive Logic Programming (ILP) problems by using an Integer Programming (IP) solver. An ILP problem is to learn a set of clauses in first-order logic from positive examples, negative examples, and background knowledge. In our method, we transform a given ILP problem to an IP problem, and then solves the IP problem by using an IP solver. We carry out experiments on classification tasks and show how our method works.