计算机应用
計算機應用
계산궤응용
COMPUTER APPLICATION
2015年
z1期
101-103,134
,共4页
王明月%王晶%齐瑞云%陈复扬
王明月%王晶%齊瑞雲%陳複颺
왕명월%왕정%제서운%진복양
智能交通系统%短时%交通流量%GMDH%预测
智能交通繫統%短時%交通流量%GMDH%預測
지능교통계통%단시%교통류량%GMDH%예측
intelligent traffic system%short-time%traffic flow%Group Method of Data Handing (GMDH)%prediction
城市交通是一个复杂的大系统,实时而准确的短时交通流量预测,可以为城市交通诱导和控制提供科学支持。针对GMDH算法建模泛化能力差的问题,结合集成学习的思想对GMDH算法进行改进,并将改进的算法应用到短时交通流量模型的构建中。结果表明,该方法可以有效地对短时交通流量进行预测,建模平均相对误差为1.10%,预测相对误差为0.58%。
城市交通是一箇複雜的大繫統,實時而準確的短時交通流量預測,可以為城市交通誘導和控製提供科學支持。針對GMDH算法建模汎化能力差的問題,結閤集成學習的思想對GMDH算法進行改進,併將改進的算法應用到短時交通流量模型的構建中。結果錶明,該方法可以有效地對短時交通流量進行預測,建模平均相對誤差為1.10%,預測相對誤差為0.58%。
성시교통시일개복잡적대계통,실시이준학적단시교통류량예측,가이위성시교통유도화공제제공과학지지。침대GMDH산법건모범화능력차적문제,결합집성학습적사상대GMDH산법진행개진,병장개진적산법응용도단시교통류량모형적구건중。결과표명,해방법가이유효지대단시교통류량진행예측,건모평균상대오차위1.10%,예측상대오차위0.58%。
The urban traffic is a complex large system, actual and accurate traffic flow prediction can provide scientific support for urban traffic guidance and control. Ensemble learning is introduced to improve the general ability of classical Group Method Of Data Handing ( GMDH ) algorithm. The short-term traffic flow model was built based on improved GMDH algorithm. Experimental results indicate that the average relative error of the model is 1. 10%, and the relative error of prediction is 0. 58%. Thus, this model is an efficient method to the short-term traffic flow forecasting.