基于PSO算法的probit模型参数估计  

Parameter Estimation of probit Model 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,嘉兴314001

出　　处：《计算机工程》2009年第23期198-200,共3页Computer Engineering

摘　　要：针对二值probit回归模型中的参数估计问题,提出一种基于粒子群优化(PSO)的参数估计算法。该算法采用以最大似然准则作为PSO的适应度函数,建立二值probit回归模型中的参数估计计算模型。数值仿真分析表明,该算法性能较好,回归结果具有较高的拟合优度。Aiming at the problem of the parameter estimation of the binary probit regression models, this paper proposes a novel algorithm to estimate parameter based on Particle Swarm Optimization（PSO） algorithm. Maximum likelihood estimation rule is adopted to be fitness function for the PSO algorithm. The model of computing parameter to the binary probit regression model is set up. Through a numerical simulation computational experiment, the effectiveness of this algorithm is demonstrated for the parameter estimation problem of the binary probit regression models. Numerical simulation analysis shows that the algorithm has better goodness of fittest in the regression result.

关 键 词：粒子群优化算法 参数估计 二值probit回归模型 最大似然估计 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]