An Improved K2 Algorithm for Symptom-Disease Bayesian Network
Ying Yu, Wenwen Yang, Hao Huang, Zheng Yang, Jianye Lu · Journal of Physics Conference Series · 2019
Abstract This paper establishes a Bayesian network for pre-diagnosis based on symptom-disease knowledge from clinical data. In order to solve the local optimal problem, the idea of simulated annealing algorithm is introduced into an improved K2 algorithm, and the BDE scoring function is used to determine the optimal network structure of the Bayesian network. performance evaluation indexes are used to evaluate the performance of the network. The experimental results show that the improved K2 algorithm outperformed the K2 algorithm.