Machine learning research project: Bayesian Network Structure Learning using Genetic Algorithms.
The experimental framework was coded in MATLAB R2016a using the Bayes Network Toolbox v1.0.3 and the Structure Learning Package v1.2.0.0.
This work presents two novel knowledge-driven parent controlling strategies for BNSL, built on top of a GA with an adaptive mutation scheme. The first improvement works upon the parent control setting, whereas the second one aims at reducing the sensitivity of such parameter by dynamically adjusting the maximum number of parents each node can hold. DOI: 10.1007/s10462-018-9615-5
This work presents a new approach which incorporates the structural properties of the BNSL problem into GA mechanisms. The proposed approach uses a new recombination operator named Parent Set crossover, capable of reducing the disruptive action of the recombination process and enhancing its exploitative power. The new operator has been implemented as part of two genetic strategies: a canonical GA and a GA with an adaptive mutation scheme. DOI: 10.1145/3071178.3071240