基于有效积温的生菜生长模型构建  

Construction of Lettuce Growth Model Based on Growing Degree Days

作　　者：段光俊 赵家松[1] 刘振洋 严伟榆[1] 马滇璟 王丽嘉 DUAN Guangjun;ZHAO Jiasong;LIU Zhenyang;YAN Weiyu;MA Dianjing;WANG Lijia(Big Data College,Yunnan Agricultural University,Kunming,Yunnan 650201;College of Data Science and Engineering,Kunming City College,Kunming,Yunnan 650032)

机构地区：[1]云南农业大学大数据学院,云南昆明650201 [2]昆明城市学院数据科学与工程学院,云南昆明650032

出　　处：《北方园艺》2024年第6期9-16,共8页Northern Horticulture

基　　金：云南省农业基础研究联合专项基金资助项目(202301BD070001-202);云南农业大学博士科研启动基金资助项目(A2032002507)。

摘　　要：以生菜为试材,采用Logistic回归、岭回归和支持向量回归3种算法,研究了不同算法对生菜生长模型的拟合效果和预测能力,以期更灵活和准确地掌握生菜的生长规律,为后续生菜生产的有效规划、生长预测和增产措施等提供参考依据。结果表明:Logistic回归、岭回归和支持向量回归的生菜生长模型实测值与预测值基于1∶1直线的平均R^(2)分别为0.818、0.897和0.957。支持向量回归对生菜生长的模拟表现最好,其次是岭回归,而Logistic回归表现最差。Taking lettuce as the test material,using three algorithms,Logistic regression,ridge regression and support vector regression,the fitting effect and prediction ability of different algorithms on the lettuce growth model were studied,in order to provide reference for effective planning,growth prediction and yield-increasing measures for subsequent lettuce production.The results showed that the average R^(2)of the measured values and predicted values of the lettuce growth model of Logistic regression,ridge regression and support vector regression based on the 1∶1 straight line were 0.818,0.897 and 0.957 respectively,and the average RMSE were 27.631,19.505 and 6.901,respectively.Therefore support vector regression performed the best in simulating lettuce growth,followed by ridge regression,and Logistic regression was the worst.

关 键 词：生长模型 LOGISTIC回归 岭回归 支持向量回归 

分 类 号：S636.2[农业科学—蔬菜学]