CAJ | 학술논문

为了分析城市公交乘客的出行特征,本文利用公交IC卡及GPS数据对公交IC卡乘客上车站点推算进行研究.针对安装车载GPS设备的车辆,运用GPS数据与IC卡数据融合算法进行推算;对于无车载GPS设备的情况,为适应一票制IC卡数据挖掘,对贝叶斯决策树算法进行改进,允许节点跳跃,推算上车站点,并且利用Markov链特性降低算法的运算复杂度.同时,本文以北京公交数据为例,对提出的两种方法进行验证.结果表明,利用本文提出的方法推算上车站点,3站之内误差的准确率达到90%以上,算法在兼顾算法精度的同时合理地控制了运算复杂度,可以实际运用于城市公交系统.
위료분석성시공교승객적출행특정,본문이용공교IC잡급GPS수거대공교IC잡승객상차참점추산진행연구.침대안장차재GPS설비적차량,운용GPS수거여IC잡수거융합산법진행추산;대우무차재GPS설비적정황,위괄응일표제IC잡수거알굴,대패협사결책수산법진행개진,윤허절점도약,추산상차참점,병차이용Markov련특성강저산법적운산복잡도.동시,본문이북경공교수거위례,대제출적량충방법진행험증.결과표명,이용본문제출적방법추산상차참점,3참지내오차적준학솔체도90%이상,산법재겸고산법정도적동시합리지공제료운산복잡도,가이실제운용우성시공교계통.
In order to analyze urban bus passengers' travel characteristic, this paper proposes several data mining algorithms for boarding stop inference based on IC card and GPS data. For those buses with GPS devices, a data-fusion method with GPS data is developed to estimate individual passenger's boarding stop. For those buses without GPS devices, an improved Bayesian decision tree algorithm with varying steps is presented to calculate the likelihood of each possible boarding stop. In addition, Markov Chain optimization technique is applied to reduce the computational complexity. Empirical data from Beijing transit route are used to validate the effectiveness of the proposed algorithms. The results demonstrate that the accuracy of identified boarding stop can be guaranteed and the algorithm complexity can be well controlled to meet the requirements of practical application. As a result, the methods can be widely adopted for urban public transportation system.