决策树学习算法在交通方式选择模型中的应用

主要阐述了决策树学习算法在交通方式选择模型中的应用.在基本决策树的基础上,使用随机森林组合学习算法来建立交通方式选择模型,以Bagging预测方法和CART算法为主,以随机特征选择和"投票"方法为辅,并相互融合,结合具体实例详细介绍该模型的建立,从数据的选择到整个森林中树的数目和每个结点处抽取的候选属性的个数调整,并对模型进行了相应的评估.实验结果表明,随机森林预测精度高,且对噪声数据具有较强的稳健性,采用决策树学习算法得出的规则在交通方式选择的分析中具有较好的实用价值.

Application of traffic mode choice model based on decision tree algorithm

The article mainly describes application research on traffic mode choice model based on decision tree algorithm.Based on the basic decision tree algorithm,traffic mode choice model based on random forest combination forecasting method is proposed.Through the integration of Bagging prediction method and the CART algorithm and random feature selection and "vote" method,the establishment of the model is described with examples from the choice of data to the entire forest tree number;and each the number of the candidate attributes extracted by the junction point adjustment;and the corresponding evaluation of the model is made.The results show that the rules obtained by decision tree method have good practical value in the analysis of the choice of traffic mode.