基于改进人工鱼群算法的回归模型参数估计

Parameters Estimation of Regression Model Based on Improved AFSA

由于人工鱼群算法在获取精确解方面存在局限性,文章对算法的参数以及人工鱼群的聚群行为进行改进,将改进的算法应用于多元线性回归模型的参数估计,并与传统的人工鱼群算法和最小二乘法的结果进行了比较,结果表明,应用改进的人工鱼群算法对多元线性回归模型做参数估计,其结果要优于人工鱼群算法和最小二乘法,为解决回归模型的参数估计问题提供了一种新的有效方法。

Due to the limitation of artificial fish swarm algorithm（AFSA） in obtaining exact solutions,this paper improves the parameters of the algorithm and the clustering behavior of artificial fish,and then applies the improved algorithm to the parameter estimation of multiple linear regression model.Finally the paper compares it with the results of traditional AFSA and least square method.It is shown that result of applying the improved ASFA to estimate the parameters of multiple linear regression model is superior to that of the artificial fish swarm algorithm and the least square method,thus providing a new effective method to solve the parameter estimation problem of regression model.