AABC算法优化ELM的心脏病辅助诊断

Cardiac assisted diagnosis based on adaptive artificial bee colony optimized ELM

针对ELM算法在心脏病辅助诊断中分类精度不高的缺陷,提出自适应人工蜂群算法优化ELM隐层输入权值和偏置的心脏病辅助诊断方法。采用自适应遗传算法对数据进行特征选择,以最优特征子集构造样本输入自适应人工蜂群算法优化ELM的分类模型。自适应人工蜂群算法改进原算法的跟随蜂概率选择机制,在搜索阶段引入最优解与次优解,通过自适应算子调整二者的引导作用。仿真结果表明,该方法相比于其它方法提高了分类精度,减少了总体耗时。

Aiming at the defect that ELM algorithm has low classification accuracy in cardiac assisted diagnosis,an adaptive arti-ficial bee colony algorithm was proposed to optimize the ELM hidden layer input weight and bias for cardiac assisted diagnosis.The adaptive genetic algorithm was used to select the features of data,and an optimal feature subset constructed sample was inputted to adaptive artificial bee colony algorithm optimized ELM classification.The adaptive artificial bee colony algorithm was used to improve the onlookers probability selection mechanism of the original algorithm.In the search phase,the optimal solution and the suboptimal solution were introduced and the guiding effects of these two were adjusted using an adaptive operator.The simulation results show that the proposed method improves the classification accuracy and reduces the overall time consumption compared to other methods.