摘要
针对粒子对算法存在过早陷入局部最优导致精度不是很高的问题,建议了一种新的基于粒子对(PPO)与极值优化(EO)混合算法。该算法利用PPO和EO的优点,借助K-means快速聚类的结果初始化其中一个粒子,并根据一定迭代次数在精英粒子对的迭代过程中引入EO算法,在保证算法收敛的同时避免后期过早陷入局部最优,从而提高聚类结果的精度。将混合算法应用于真实的基因表达数据。实验结果表明,混合算法比K-means和粒子对算法具有更好的聚类精度和稳定性。
In order to solve the problem that particle pair algorithm exists local optimization premature to lower precision,this paper suggested a new hybrid algorithm based on particle pair optimization（PPO） and extremal optimization（EO）.The hybrid algorithm used the merits of PPO and EO,and assigned the fast cluster result of the K-means to initialize a particle and introduced the extremal optimization algorithm in the iteration process of elitist particle pair according to interval iteration,which could ensure convergence and avoid local optimization premature in the later period,so it improved the precision of the clustering result.Applying the hybrid algorithm to gene expression data,the experiment results indicate that the hybrid algorithm obtains better clustering precision and stability than the K-means algorithm and particle pair algorithm.
出处
《计算机应用研究》
CSCD
北大核心
2011年第10期3675-3677,3680,共4页
Application Research of Computers
关键词
基因聚类
K-MEANS算法
粒子对
极值优化算法
混合算法
gene clustering
K-means algorithm
particle pair
extremal optimization algorithm
hybrid algorithm
