Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic reasoning in intelligent systems: networks of plausible inference
Bayesian inductive logic programming
COLT '94 Proceedings of the seventh annual conference on Computational learning theory
Inductive logic programming: derivations, successes and shortcomings
ACM SIGART Bulletin
Data mining in finance: advances in relational and hybrid methods
Data mining in finance: advances in relational and hybrid methods
Bayesian Networks and Decision Graphs
Bayesian Networks and Decision Graphs
Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control and Artificial Intelligence
Algorithmic Program DeBugging
Probabilistic Networks and Expert Systems
Probabilistic Networks and Expert Systems
Learning Logical Definitions from Relations
Machine Learning
Learning Nonrecursive Definitions of Relations with LINUS
EWSL '91 Proceedings of the European Working Session on Machine Learning
ACM SIGKDD Explorations Newsletter
Learning fuzzy rules with their implication operators
Data & Knowledge Engineering
Mining associative classification rules with stock trading data - A GA-based method
Knowledge-Based Systems
Fitness function based on binding and recall rate for genetic inductive logic programming
ICSI'12 Proceedings of the Third international conference on Advances in Swarm Intelligence - Volume Part I
A new relational Tri-training system with adaptive data editing for inductive logic programming
Knowledge-Based Systems
| Hi-index | 12.05 |
Inductive logic programming (ILP) is one of the most popular approaches in machine learning. ILP can be used to automate the construction of first-order definite clause theories from examples and background knowledge. Although ILP has been successfully applied in various domains, it has the following weaknesses: (1) weak capabilities in numerical data processing, (2) zero noise tolerance, and (3) unsatisfactory learning time with a large number of arguments in the relation. This paper proposes a phenotypic genetic algorithm (PGA) to overcome these weaknesses. To strengthen the numerical data processing capabilities, a multiple level encoding structure is used that can represent three different types of relationships between two numerical data. To tolerate noise, PGA's goal of finding perfect rules is changed to finding top-k rules, which allows noise in the induction process. Finally, to shorten learning time, we incorporate the semantic roles constraint into PGA, reducing search space and preventing the discovery of infeasible rules. Stock trading data from Yahoo! Finance Online is used in our experiments. The results indicate that the PGA algorithm can find interesting trading rules from real data.