%0 Journal Article
%T 基于回归方法的鲍鱼年龄预测<br>Prediction of Abalone Age Based on Regression Methods
%A 孙丽娜
%J Statistics and Applications
%P 1485-1498
%@ 2325-226X
%D 2023
%I Hans Publishing
%R 10.12677/SA.2023.126152
%X 本文基于物理测量确定鲍鱼年龄的方法，根据测量数据，利用R语言，建立线性回归、逻辑回归、岭回归、LASSO回归模型，来预测鲍鱼的年龄。并通过平均绝对误差MAE、均方误差MSE和对称平均绝对百分比误差SMAPE对模型进行评价，结果表明，LASSO回归模型的拟合优度更好。考虑到变量间相关性强，可能存在多重共线性，本文利用偏最小二乘及主成分分析两种方法对变量降维，降维后再进行回归分析，以期消除多重共线性对模型带来的影响。利用MSE评价模型，结果表明，这两种降维方法都没能减小MSE，反而得到模型的MSE更大。<br />
In this paper, based on physical measurements to determine the age of abalone, linear regression, logistic regression, ridge regression, and LASSO regression models are established to predict the age of abalone based on the measurement data, using R language. The models are evaluated by mean absolute error MAE, mean square error MSE, symmetric mean absolute percentage error SMAPE, and the results show that the LASSO regression model has a better goodness of fit. Consid-ering the strong correlation between the variables and the possible existence of multicollinearity, this paper uses two methods of partial least squares and principal component analysis to reduce the dimensionality of the variables, and then regression is performed after the reduction of dimen-sionality, in order to eliminate the impact of multicollinearity on the model. Using MSE to evaluate the model, the results show that both methods of dimensionality reduction fail to reduce the MSE, but instead, the MSE of the model is obtained to be larger.
%K 线性回归，逻辑回归，岭回归，LASSO回归，偏最小二乘，主成分分析<br>Linear Regression
%K Logistic Regression
%K Ridge Regression
%K LASSO Regression
%K Partial Least Squares
%K Principal Component Analysis
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=76612