计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (23): 112-118.

• 数据库、数据挖掘、机器学习 • 上一篇    下一篇

Ant-FOIL:一种归纳逻辑蚁群学习系统

颜晨阳   

  1. 宁波大学 职教学院,浙江 宁波 315000
  • 出版日期:2015-12-01 发布日期:2015-12-14

Ant-FOIL:ant colony inductive logical learning system

YAN Chenyang   

  1. School of Vocational Tech., Ningbo University, Ningbo, Zhejiang 315000, China
  • Online:2015-12-01 Published:2015-12-14

摘要: 提出了一种新型的归纳逻辑学习系统Ant-FOIL,该系统结合了蚁群优化元启发方法和FOIL算法,利用蚁群系统优化框架来指导FOIL进行规则构建。系统在若干经典归纳逻辑测试数据集上进行了测试,结果显示Ant-FOIL在预测准确率等关键指标上的表现要优于标准FOIL算法以及FOIL系统的若干改进变体(mFOIL、nFOIL、kFOIL和GILP)和其他一些成熟的归纳逻辑学习算法(Aleph和FOIL)。

关键词: 归纳逻辑编程, 一阶逻辑, 蚁群系统

Abstract: A novel inductive logical learning system Ant-FOIL is presented. It tightly integrates the ant colony optimization meta-heuristic method with the inductive logic programming rule-learner FOIL. Ant-FOIL employs the Ant Colony System (ACS) scheme to directly guide its rule construction. Experiments in applying Ant-FOIL to well-known benchmarks show that Ant-FOIL performs better than either its base line algorithm FOIL or certain FOIL variants(mFOIL, kFOIL, nFOIL and GILP), and is at the same time competitive with more sophisticated approaches(Aleph and FOIL).

Key words: inductive logical programming, first order logic, ant colony system