The structure learning of Bayesian networks is a NP-hard problem, which cannot be easily solved since it is usually a complex combination optimization problem. Thus, many structure learning algorithms using evolutionary techniques are investigated recently to obtain a reasonable result. However, evolutionary algorithms may suffer from a low accuracy and restricts their applications. In this paper, we apply the Biased Random-Key Genetic Algorithm to solve Bayesian network structure learning problem since this framework is novely designed to solve conventional combination optimization problems. Also, we use a local optimization algorithm as its decoder to improve the performance. Experiments show that our method achieves better performances on the real-world networks than other state-of-art algorithms.
CITATION STYLE
Sun, B., & Zhou, Y. (2021). Biased Random-Key Genetic Algorithm for Structure Learning. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 12689 LNCS, pp. 399–411). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-78743-1_36
Mendeley helps you to discover research relevant for your work.