A Novel Particle Swarm Optimization Approach for Patient Clustering From Emergency Departments
A Novel Particle Swarm Optimization Approach for Patient Clustering From Emergency Departments
复制标题
DOI：
10.1109/tevc.2018.2878536
复制
发表时间：
2019-08-01
影响因子：
14.3
通讯作者：
Bell, David
中科院分区：
文献类型：
--
作者：
Liu, Weibo;Wang, Zidong;Bell, David
在本文中，提出了一种新颖的粒子群优化（PSO）算法，以提高传统聚类方法的准确性，并应用于分析英国当地医院事故与急诊（A＆E）部门的实时患者出勤数据。在所提出的随机分布延迟 PSO（RODDPSO）算法中，根据速度更新模型是否从一种模式切换到另一种模式，通过评估每次迭代中的进化因子来确定进化状态。为了减少陷入局部最优的可能性并扩大搜索空间，以分布式方式在速度更新模型中引入了反映先前个人最佳和全局最佳粒子历史的随机发生的时间延迟。采用八个著名的基准函数来评估所提出的 RODDPSO 算法，通过广泛的比较表明该算法优于一些当前流行的 PSO 算法。为了进一步说明其应用潜力，RODDPSO 算法成功地应用于西伦敦当地急症室的患者聚类问题中进行数据分析。实验结果表明，基于RODDPSO的聚类方法优于其他两种著名的聚类算法。
In this paper, a novel particle swarm optimization (PSO) algorithm is proposed in order to improve the accuracy of traditional clustering approaches with applications in analyzing real-time patient attendance data from an accident & emergency (A&E) department in a local U.K. hospital. In the proposed randomly occurring distributedly delayed PSO (RODDPSO) algorithm, the evolutionary state is determined by evaluating the evolutionary factor in each iteration, based on whether the velocity updating model switches from one mode to another. With the purpose of reducing the possibility of getting trapped in the local optima and also expanding the search space, randomly occurring time-delays that reflect the history of previous personal best and global best particles are introduced in the velocity updating model in a distributed manner. Eight well-known benchmark functions are employed to evaluate the proposed RODDPSO algorithm which is shown via extensive comparisons to outperform some currently popular PSO algorithms. To further illustrate the application potential, the RODDPSO algorithm is successfully exploited in the patient clustering problem for data analysis with respect to a local A&E department in West London. Experiment results demonstrate that the RODDPSO-based clustering method is superior over two other well-known clustering algorithms.