摘要
针对城市居民是否选择低碳出行的问题,本文基于机器学习方法,通过问卷调查的形式,对影响城市居民低碳出行的相关因素与实际交通出行方式进行数据收集,设置自变量与因变量及其取值,并基于Logistic回归模型原理进行拟合建模,在不同属性下,对城市居民是否选择低碳出行方式和影响进行分析,得到不同属性之间与各个属性内部的影响与区别关系。同时,运用Logistic回归、神经网络与决策树3种分类预测模型,在现有数据基础上,对城市居民是否选择低碳出行方式进行预测。研究结果表明,年龄、收入、距离和是否拥有汽车及驾驶证,对选择低碳出行方式具有重要影响,尤其是没有汽车的居民,选择低碳出行方式的可能性是拥有汽车居民的9.065倍,且3种模型的预测准确度均在85%以上,而神经网络模型的预测准确度接近90%,各项指标均说明预测效果良好,验证了该方法的可行性及适用性。该研究为预测不同地区与类型的低碳出行方式提供了理论参考。
In response to the issue of whether urban residents choose low-carbon travel,this article provides a study based on machine learning methods.Through a questionnaire survey,data is collected on the relevant factors and actual transportation modes that affect urban residents’low-carbon travel.Independent variables and dependent variables are set,and their values are set.Based on the principle of logistic regression model,fitting modeling is carried out to analyze whether urban residents choose low-carbon travel modes and their impacts under different attributes,Obtain the influence and differential relationship between different attributes and within each attribute.At the same time,three classification and prediction models,logistic regression,neural network,and decision tree,are used to predict whether urban residents choose low-carbon travel modes based on existing data.The research results indicate that age,income,distance,and ownership of a car and driver’s license have important impacts on choosing low-carbon travel methods,especially for residents without a car.The likelihood of choosing low-carbon travel methods is 9.065 times higher than that of residents with a car,and the prediction accuracy of the three models is above 85%,while the prediction accuracy of the neural network model is close to 90%.All indicators indicate that the prediction effect is good,the feasibility and applicability of this method have been verified.This study provides a theoretical reference for predicting low-carbon travel modes in different regions and types.
作者
许佳佳
陈慧敏
XU Jiajia;CHEN Huimin(School of Traffic Engineering,Anhui Sanlian University,Hefei 230601,China;School of Politics and Public Administration,Guangxi Minzu University,Nanning 530006,China)
出处
《青岛大学学报（工程技术版）》
CAS
2023年第3期30-38,共9页
Journal of Qingdao University(Engineering & Technology Edition)
基金
安徽省高校自然科学优秀青年科研项目(2022AH030161)
安徽省高校自然科学重点科研项目(2022AH051993)
安徽三联学院科研项目(PTZD2022008,zjt22002)。
