针对单一聚类算法存在的不能泛化的问题,将集成学习技术应用于聚类算法中,集成学习技术可以显著提高学习系统的泛化能力。提出了1种基于粒子群和遗传算法的协同进化聚类集成算法,粒子群算法保证算法快速收敛,遗传算法全局搜索扩大搜索范围,提高了聚类的性能和收敛速度。将本研究提出的算法在多个UCI数据集上进行试验验证,结果表明该算法是有效的。
Since clustering could not solve the problem of generalization, the integration technology was introduced into clustering algorithm, which could significantly improve the generalization ability of learning systems. A co-evolutionary clustering ensemble algorithm based on particle swarm optimization and genetic algorithm (CEGPCE) was proposed. PSO (particle swarm optimization) ensured the algorithm with fast convergence, and GA (genetic algorithm) expanded the search scope with its global search capability, which improved the performance of the algorithm and the convergence speed. Experiments on the UCI data sets verified the effectiveness of CEGPCE.