基于改进麻雀搜索算法的优化型极限学习机

An Optimized Extreme Learning Machine Based on Improved Sparrow Search Algorithm

为提升极限学习机的性能,文章利用改进的麻雀搜索算法对极限学习机的参数进行优化。首先,提出一种菱形分组机制用于增加算法种群多样性和融合模拟退火思想改善算法陷入局部极值点的缺陷。其次,通过10个基准函数进行仿真测试,实验结果表明,改进的麻雀搜索算法在大部分测试函数上表现出更好的性能。最后,将改进的算法用于优化极限学习机的输入权阈值,通过基准数据集仿真测试,优化后的极限学习机在建模精度上平均提高了7.4%。

In order to improve the performance of the Extreme Learning Machine(ELM),this paper proposes to optimize the parameters of the ELM with an Improved Sparrow Search Algorithm(ISSA).Firstly,a rhombus grouping mechanism is proposed to increase the diversity of the algorithm population and to improve the defect of the algorithm trapped in local extreme points by incorporating the Simulated Annealing idea.Then,the simulation test is carried out through 10 benchmark functions,and the experimental results show that ISAA has better performance on most test functions.Finally,ISSA is used to optimize the input weight threshold of the ELM.Through the simulation test of the benchmark data set,the optimized ELM has an average improvement of 7.4%in modeling accuracy.