出租车乘车概率预测的置信规则库推理方法.pdfVIP

ISSN l673．9418 C0DEN JKYTA8 E—mail：fcst@vip．163．com JoumalofFrontiersofComputerSciencemadTechnology http：／／www．ceaj．org 1673-·9418／2015／09(08)·0·985-·10 Tel：+86．．10． doi：10．3778~．issn．1673—9418．1409066 出租车乘车概率预测的置信规则库推理方法 杨隆浩，蔡芷铃 ，黄志鑫，何 星，傅仰耿 福州大学 数学与计算机科学学院，福州 350116 BeliefRule—BaseInferenceM ethodologyforPredictingProbabilityofTakingTaxi YANG Longhao，CAIZhiling，HUANGZhixin，HEXing，FUYanggeng CollegeofM athematicsandComputerScience，FuzhouUniversity，Fuzhou350116，China +Correspondingauthor：E—mail：ygfu@qq．com YANG Longhao，CAIZhiling，HUANG Zhixin，eta1．Beliefrule-baseinferencemethodology forpredicting probabilityoftakingtaxi．JournalofFrontiersofComputerScienceandTechnology,2015，9(8)：985—994． Abstract：Largescaleofdata，varioustypesoflow—levelattributesanduncertainytofpredictioninformationexistin probabiliytpredictionoftakingtaxi．Tosolvetheseproblems，thispaperoffiinedealswiththeGPSdataoftaxiand roadnetworkdatabyusingminingalgorithmsinthelarge-scaletrajectorydatadomain，thenbuildsabelief rule—basebytransformingvariousytpesofinformationwithuncertaintyintourleswhichareinfomr ofthebelief structure，afterthatusesRIMER(beliefurle—baseinferencemethodologyusingevidentialreasoning)togetthefinal probabiliytofanypointsontheroadnetwork．Finally,theGPSdataofBeijing’StaxiinNovemberof2012are takenasanexampletoillustratetheusageoftheonlinepredictionmethod，andhteresultsshow thereal—timeand accuracyoftheproposedmethod． Keywords：probabiliytprediction；GPSdata；roadnewtorkdata；beliefurle--base；beliefrule-·baseinferencemethod。- ologyusingevidentialreasoning(RIMER) 摘 要：出租车乘车概率预测中存在数据量级大，底层属性类型多，预测信息不确定的问题。鉴于此，整合大 规模轨迹数据范畴中现有的挖掘算法对出租车GPS数据和路网数据进行离线处理；将多类型的不确定性数据 TheNationalNaturalScienceFoun