基于Lasso回归的北京地区黄瓜价格波动原因分析

Cause Analysis of Cucumber Price Fluctuation in Beijing Based on Lasso Regression

近年来蔬菜市场价格波动幅度较大,为探究蔬菜价格波动剧烈的原因,以北京地区黄瓜为例,对2010—2017年价格波动规律进行实证分析基础上探究了Lasso回归方法的适用性,并使用Lasso回归方法对黄瓜数据进行建模并求解。该方法剔除了相关度较小的因素,获取主要影响因素及其相关度。相较于常用的最小二乘法求解,多重共线性判定条件值仅有19.66,拟合系数达0.844 8,证明Lasso回归模型适用于蔬菜价格原因分析且相比于传统方法有更好的表现,可为进一步开展蔬菜价格的预测提供参考依据。

In recent years,the price of vegetables had fluctuated greatly.The study used cucumber in Beijing to explor the applicability of Lasso regression method based on empirical analysis of price fluctuations from 2010 to 2017,in order to explore the causes of sharp fluctuations in vegetable prices.Used Lasso regression method to model and solve cucumber data.The method eliminated the factors with small correlation degree and obtained the main influencing factors and their correlation degree.Compared with the commonly used least square method,the multicollinearity determination condition value was only 19.66,and the fitting coefficient reached 0.844 8.The result proved that Lasso regression model was suitable for vegetable price cause analysis and had better performance than traditional methods.This study could provide basis for further vegetable price prediction.