This is a repository for university student innovation project.
This project has designed a novel discretized particle swarm optimization algorithm, termed Aprori-DPSO, for structural learning of Bayesian networks. This algorithm employs strong association rules derived from the Apriori algorithm as prior knowledge. The movement velocities of particles are calculated using matrix operations. Through a mapping function, the distribution of particle velocities is adjusted to facilitate the discretization of velocity values. Subsequently, the positions are updated based on the discrete velocity matrix, enabling the learning of network structures that more accurately reflect real-world scenarios. This model is utilized for predicting the severity of accidents.