论文部分内容阅读
个性诊断(Personality Diagnosis,PD)算法只适合应用于离散型空间,应用到连续型空间时需要先进行离散化处理,导致其应用范围受限。对个性诊断算法进行改进,并应用到实时路况估计中,提出基于连续型的个性诊断和热门路段的路况估计模型(Improved Personality Diagnosis and Popular Road Model,IPDPR)。首先根据数据覆盖率提取高覆盖率路段作为基准个性类型;然后判别路网中所有路段的类型,即计算相似性概率;接着根据相似性概率获得缺失项的取值概率分布;最后计算概率最大值作为估计值。实验结果表明,本文所提IPDPR模型估计误差比概率主成分分析(Probabilistic Principal component analysis,PPCA)算法小53.88%,比滑动平均法小11.47%.
The personality diagnosis (Personality Diagnosis, PD) algorithm is only suitable for discrete space. When applied to continuous space, the discrete algorithm needs to be discretized so that its application range is limited. In this paper, the personality diagnosis algorithm is improved and applied to the real-time traffic estimation. The improved Personality Diagnosis and Popular Road Model (IPDPR) is proposed based on continuous personality diagnosis and hot section estimation. First of all, according to the data coverage, the high coverage section is extracted as the baseline personality type; then the type of all road sections in the road network is identified, that is, the similarity probability is calculated; then the probability distribution of missing items is obtained according to the similarity probability; finally, the maximum probability As an estimate. Experimental results show that the estimated error of the proposed IPDPR model is 53.88% less than that of the Probabilistic Principal Component Analysis (PPCA) algorithm, which is 11.47% less than the sliding average method.