基于混合蛙跳算法的高维生物医学数据特征选择方法

针对高维生物医学数据包含大量无关或弱相关特征,影响疾病诊断效率的现状,提出了一种基于改进混合蛙跳算法的高维生物医学数据特征选择方法。该方法将混沌记忆权重因子和平衡分组策略引入基本混合蛙跳算法,在强化算法多样性的同时,维持了算法全局和局部寻优之间的平衡,降低了算法陷入局部最优的可能,进一步提高了混合蛙跳算法特征选择方法在特征空间的探索能力。实验结果表明:与改进遗传算法、粒子群优化算法特征选择方法比较,改进混合蛙跳算法特征选择方法在高维生物医学数据特征子集识别、分类精度方面取得了更好的效果。

Feature selection of high-dimensional biomedical data based on shuffled frog leaping algorithm

Concerning the high-dimensional biomedical data contains a large number of irrelevant or weak correlation features,affecting the efficiency of disease diagnosis,this paper proposed a high-dimensional biomedical data feature selection method based on improved shuffled frog leaping algorithm.This method introduced chaos memory weighting factor and balance group strategy into the shuffled frog leaping algorithm.The method enhanced the diversity of the algorithm,maintained the balance between global and local optimization of the algorithm,reduced the probability of the algorithm falling into local optimality and further improved the exploring ability of the shuffled frog leaping algorithm in feature space.The simulation results show that compared with the improved GA and PSO algorithm,the improved shuffled frog leaping algorithm feature selection method can significantly enhance the recognition and classification accuracy of related feature subsets.