Rule Induction Using Set-Based Particle Swarm Optimisation
Jean-Pierre van Zyl, Andries Petrus Engelbrecht · 2022 IEEE Congress on Evolutionary Computation (CEC) · 2022
This paper presents a new approach to induce a list of rules from a dataset by using a set-based particle swarm optimisation algorithm. Many contemporary rule induction algorithms tend to use similar information gain based approaches to fit a training dataset. The proposed novel algorithm is a meta-heuristic approach which finds an optimal rule list while providing the flexibility to overcome traditional drawbacks such as overfitting and rigidity to the datatypes that can be used. This paper shows that the proposed algorithm performs comparatively well when compared to existing rule induction algorithms and it has the potential to be expanded further by adding rule pruning techniques.