Bayesian Network Structure Learning Based on Small Sample Data

Abstract

Aiming at the problem that it is difficult to learn the optimal solution of Bayesian structure under the condition of small sample data, this paper proposes a Bayesian structure learning algorithm under small data set. Firstly, an improved Bootstrap sampling is proposed to expand the small data, and the extended sample is modified through the maximum weight spanning tree. Secondly, the standard particle swarm optimization (PSO) algorithm is improved, and the calculation method in the update formula is redefined to adapt to Bayesian network structure learning. Finally, the simulation verification of a calculation example proves the effectiveness of the improved algorithm for Bayesian network structure learning.