基于改进MPA优化的高斯混合模型算法

Gaussian Mixture Model Algorithm based on Improved MPA Optimization

摘要:

针对高斯混合模型算法（GMM）对初始参数敏感、易陷入局部最优的问题，本文提出一种基于改进海洋捕食者算法优化的GMM算法（MMPA-GMM）。首先基于混沌序列和伪对立学习策略初始化种群，引入非线性收敛因子平衡MPA算法的全局与局部搜索，同时提出融入社会等级制度的位置更新策略；然后从搜索能力和收敛速度对改进的MPA进行分析；最后以S_Dbw指标作为算法的适应度函数，利用改进的MPA优化GMM算法的初始参数。实验结果表明，改进的MPA在4种测试函数上表现良好，并且MMPA-GMM算法对4个数据集的聚类效果均有改善，有效避免了GMM算法陷入局部最优的问题。

Abstract:

Aiming at the problem that the Gaussian Mixture Model Algorithm is sensitive to initial parameters and easy to fall into local optimality, a GMM algorithm based on improved marine predator algorithm optimization is proposed. First, initialize the population based on the chaotic sequence and the pseudo-opposition learning strategy. Secondly, introduce a nonlinear convergence factor to balance the global and local search of the MPA algorithm, and propose a location update strategy integrated into the social hierarchy. Then, analyze the improved MPA from the searchability and convergence speed. Finally, the S_Dbw index is used as the fitness function of the algorithm, and the improved MPA is used to optimize the initial parameters of the GMM algorithm. The experimental results show that the improved MPA performs well on the five test functions, and the MMPA-GMM algorithm improves the clustering effect of the four data sets, effectively avoiding the problem of the GMM algorithm falling into the local optimum.