基于数据挖掘的市民出行公交线路预测研究

Research on Public Transit Route Forecast Based on Data Mining

近年来一些学者大多利用海量公交卡信息对整体客流进行预测,并无针对具体乘客在公交线路上的出行做出预测。基于数据挖掘技术,以广东省部分公交线路上的历史公交卡交易数据为基础,采用滑动窗口模型构造训练样本和测试样本,借助随机森林算法,对市民在公交线路上的出行进行预测研究,并用精度、召回率以及F1值进行评估,证明模型的可行性。

In recent years, some scholars mostly use mass transit card information to forecast the overall passenger flow, but there is no prediction about specific passenger who travels in the bus line. Based on the data mining technology, explores and predicts the travel patterns of the citizens on the basis of historical bus card transaction data of some bus lines in Guangdong Province, and uses the sliding window model to construct the training and test samples. Moreover, uses random forest algorithm, proves the feasibility of the model by the precision, recall rate and F1 value.