引用本文:
  • 张贺,范梦萱.基于Lasso回归模型的青岛海洋经济和海洋产业分析[J].海洋开发与管理,2022,39(8):22-28    
【打印本页】 【下载PDF全文】 查看/发表评论关闭

←前一篇|后一篇→

过刊浏览    高级检索

本文已被:浏览 660次   下载 390 本文二维码信息
码上扫一扫!
基于Lasso回归模型的青岛海洋经济和海洋产业分析
张贺,范梦萱
中国海洋大学
摘要:
为促进青岛海洋产业结构优化和海洋经济高质量发展,文章构建评价指标体系,采用Lasso回归模型分析对青岛海洋经济发展产生重要影响的海洋产业,并提出发展建议。研究结果表明:青岛海洋经济发展水平评价指标体系包括海洋渔业、海洋化工业、海洋进出口总额、海洋环境、海洋交通运输业和滨海旅游业6个大类和20个特征变量;采用10折交叉验证确定调节系数的最优取值,并通过求解Lasso回归系数最终筛选10个特征变量进入回归模型;根据回归系数估计值,国内旅客数量的重要性最强,海洋渔业产值的重要性较强,而货物吞吐量的重要性较弱;Lasso回归模型的预测准确率、模型泛化能力和可信度均较高,属于稳健的稀疏模型;未来青岛应进一步发挥滨海旅游业和海洋渔业等的产业优势,着力发展海洋交通运输业,保护海洋生态环境以及加大海洋科技研发投入。
关键词:  海洋产业  海洋经济  交叉验证  模型泛化  调节系数
DOI:10.20016/j.cnki.hykfygl.20220803.002
投稿时间:2021-12-04修订日期:2022-07-22
基金项目:教育部人文社会科学规划基金项目“基于高维高频复杂数据的非等间隔日内市场风险评估与应用研究”(18YJA790113).
The Marine Economy and Marine Industry in Qingdao Based on Lasso Regression Model
ZHANG He,FAN Mengxuan
Ocean University of China
Abstract:
In order to promote the optimization of marine industry structure and the high-quality development of marine economy in Qingdao, this paper established the evaluation index system, analyzed the marine industry which had an important impact on the development of marine economy in Qingdao by using Lasso regression model, and put forward development suggestions. The results showed that the evaluation index system of Qingdao′s marine economic development level included 6 categories and 20 characteristic variables, which were marine fishery, marine chemical industry, total marine import and export, marine environment, marine transportation and coastal tourism. The optimal value of the adjustment coefficient was determined by 10 fold cross validation, and 10 characteristic variables were selected into the regression model by solving the Lasso regression coefficient. According to the estimated value of regression coefficient, the number of domestic passengers was the most important, the value of marine fishery output was important, and the value of cargo throughput was the least important. Lasso regression model had high prediction accuracy, model generalization ability and reliability, and belonged to robust sparse model. In the future, Qingdao should give full play to the industrial advantages of coastal tourism and marine fishery, focus on the development of marine transportation industry, protect the marine ecological environment and increase the investment in marine science and technology research and development.
Key words:  Marine industry,Marine economy,Cross validation,Model generalization,Adjustment coefficient