Bayesian Network Structure Discovery Using Antlion Optimization Algorithm

Hoshang Qasim Awla ； Shahab Wahhab Kareem ； Amin Salih Mohammed

《International Journal of Systematic Innovation》 7卷1期 (2022/03) Pp. 46-65

https://doi.org/10.6977/IJoSI.202203_7(1).0004

摘要

Bayesian networks have recently been used for discovering an optimal learning structure in machine learning. Bayes networks can describe possible dependencies of explanatory variables. As a novel approach to studying the structure of a Bayesian network, the authors present the Antlion Optimization Algorithm (ALO). In the algorithm; deletion, rewind, insertion, and change are utilized to produce ALO to reach the best hull solution. Essentially, the technique used in the ALO algorithm imitates the antlions' behaviors while hunting. The suggested approach is contrasted with simulated Annealing, Simulated Annealing Hybrid Bee, Greedy Search Hybrid Bee, optimization inspired by Pigeon, and greedy search using the BDe Score function. The researchers also studied the representation of the confusion matrix of these techniques using different reference data sets. The findings of the assessments reveal that the proposed algorithm works better than the other algorithms and has better consistency and score values. As shown by the experimental evaluations, the proposed method has a more reliable performance than other algorithms (including the production of excellent scores and accuracy values).

關鍵字

Bayesian Network ； Metaheuristics ； Antlion Optimization ； Structure Learning ； Hunting Search ； Local & Global Search ； Search and Score