基于粒子群算法的Logistic回归模型参数估计  

Parameter estimation method of logistic regression models based on particle swarm optimization algorithm

作　　者：刘锦萍[1,2] 郁金祥[2] 

机构地区：[1]华东师范大学计算机科学系,上海200062 [2]嘉兴学院数学与信息工程学院,浙江嘉兴314001

出　　处：《计算机工程与应用》2009年第33期42-44,47,共4页Computer Engineering and Applications

摘　　要：针对Logistic回归模型中的参数估计计算复杂难题,提出一种基于粒子群优化算法(PSO)的估计方法。以最大似然准则作为粒子群优化算法的适应度函数,建立了Logistic回归模型中的参数估算模型。数值仿真分析表明,粒子群优化算法可以更精确地计算出相关参数。In order to solve the computing complex problems of the parameter estimation to the logistic regression models,a novel method to estimate parameter is presented based on particle swarm optimization algorithm.Maximum likelihood estimation is adopted to be fitness function for the optimization problem.Thus the model of computing parameter to the logistic regression models is set up.The numerical simulation results show that PSO algorithm can be used to calculate the parameters of the logistic regression models.

关 键 词：粒子群优化算法 参数估计 LOGISTIC回归模型 最大似然估计 

分 类 号：O211.3[理学—概率论与数理统计] TP301.6[理学—数学]