基于决策树的交通拥挤状态预测

Prediction of Traffic Congestion Based on Decision Tree

随着交通需求量的增加,城市道路拥挤现象越来越严重,所造成的直接或间接的经济损失是难以估量的.若能可靠地预测出即将形成的交通拥挤状态,并采取及时、有效的交通管理措施,可以避免交通拥堵的产生或减轻其严重程度.城市智能交通信息平台积累了大量的交通数据,各交通数据往往存在某些内在的联系.分析影响交通拥挤状态的环境因素,利用数据挖掘中的决策树方法对大量已有历史交通数据进行挖掘可以确立交通拥挤发生模式,从而利用发生模式与当前数据来预测交通流拥挤状态.决策树擅长处理非数值数据,而且计算速度快,对交通拥挤状态预测具有较强的适用性.

With the increase of traffic demend,urban traffic congestion becomes more serious.The direct or indirect economic loss caused by traffic congestion is amazing.Managers can take measures timely if the forthcoming congestion can be reliably predicted.Urban traffic information platform has accumulated a large number of traffic data and the traffic data often intrinsically linked to each other.The data mining method of decision tree can be used to analysis the historical data and find the congestion patterns,then predict the current state of traffic congestion.Decision tree is good at dealing with non-numerical value and it can calculate rapidly,so it is applicable for the prediction of congestion state.