一种基于多源数据的出租车分布预测方法研究

A Taxi Travel Forecasting Method Based on Multi-source Data

为了改善传统的交通需求预测方法以居民出行OD调查为基础,得出的交通分布结果受样本量等因素影响,预测值与实际值相差较大的问题,本文提出基于现有出租车GPS数据、计价器数据和电召数据的一种基于多源数据融合的出租车分布预测方法,进行出租车出行分布预测.该方法可以根据多源历史数据估计出租车OD分布,并可通过预测未来出租车OD分布,提高预测准确率.

The traditional method of travel forecasting based on the analysis of residents travel OD survey,is often limited due to the sample size and other factors. This paper proposes a travel prediction method based on the existing taxi GPS data,pricing data,and phone-call data. The method can estimate the distribution status of the city＇s taxi traffic based on historical data and can predict the distribution of taxi traffic in planning year. The forecasting accuracy rate of taxi trip distribution will be greatly enhanced.