摘要： 针对传统公交车到站时间预测算法精度较低的问题，提出一种利用条件映射进行时间预测的模型。在兼顾拟合度和泛化能力的同时，模型直接将交通信息映射为到站所需时间。采用基于范例的推理技术，给出条件映射预测模型的实现算法，并用实测数据对预测算法进行比较。实验结果表明，在正常时段和高峰期，该算法的预测精度分别为100%和85%，平均误差分别为13 s和30.5 s。
关键词:
智能公交,
到站时间,
预测算法,
条件映射,
基于范例的推理
Abstract: In view of the shortcomings of existing Bus Arrival Time Prediction(BATP) model, this paper proposes a novel BATP model based on condition mapping. The model maps traffic information to arrival time directly, which makes it have the advantage of simpleness, high degree of fitting and strong generalization ability. It proposes an algorithm for the model based on Case-based Reasoning(CBR), and testes the algorithms with actual data. Experiments in the peak and normal times show that the prediction accuracy of the algorithm is 100% and 85%, and the average deviation is 13 seconds and 30.5 seconds respectively.
Key words:
intelligent public transportation,
arrival time,
prediction algorithm,
condition mapping,
Case-based Reasoning(CBR)
中图分类号:
陈圣兵, 李正茂, 王晓峰. 智能公交中基于条件映射的到站时间预测算法[J]. 计算机工程, 2012, 38(11): 147-149,152.
CHEN Ku-Bing, LI Zheng-Mao, WANG Xiao-Feng. Arrival Time Prediction Algorithm for Intelligent Public Transportation Based on Condition Mapping[J]. Computer Engineering, 2012, 38(11): 147-149,152.