基于樽海鞘算法优化的帕金森病早期诊断模型研究与并行优化

帕金森病是一种常见的神经性慢性疾病,由于其病因尚不明确,导致早期诊断精度低的问题,提出一种改进的优化核极限学习机方法用于帕金森病的早期诊断.研究利用混沌理论和高斯变异方法改进樽海鞘算法(salp swarm algorithm,SSA),提出一种基于进化机制的智能诊断模型ISSA-KELM.改进的SSA算法同步实现特征选择和KELM核函数的参数优化,有效地解决了模型的参数设定和最优特征选择问题,并基于OpenMP平台多线程调度处理模型,在保证模型分类精度最大化的同时进一步提高计算效率.实验结果表明,提出模型在分类精度上高于已有方法,计算效率也得到极大提高,具有较好的综合性能,验证了本模型有着很好的应用前景,有助于辅助临床医生在诊断中作出更准确的决策.

Research and parallel optimization of Parkinson's disease early diagnosis model based on improved salp swarm algorithm

Parkinson''s disease(PD) is a common chronic neurological disease. Because of its unclear etiology, the early diagnosis of is quite difficult. This article proposed a novel model based on improved optimized kernel extreme learning machine(KELM) for the early diagnosis of PD. This paper used chaos theory and Gauss mutation factor to improve salp swarm algorithm(SSA), and designed the intelligent diagnosis model named as ISSA-KELM based on evolution mechanism. The improved SSA(ISSA) used to conduct adjust KELM important parameters and feature subsets selection synchronously. Moreover, the model was processing on multi-thread scheduling on openMP platform, which could achieve the maximum classification accuracy of model, and further improved the computational efficiency. The experimental results show that the classification accuracy of the proposed model is higher than those of the existing methods, the computational efficiency is algo greatly enhanced, it demonstrated the good comprehensive performance. Therefore, it proves that the proposed model has goods prospects for application and it will be helpful for clinicians to make more accurate decision in diagnosis of PD.