【摘要】 由于共享单车的流动性强,随机性很高,因此快速精确地预测出城市共享单车的短时需求量具有十分重要的意义。采用随机森林、极端随机树、支持向量机、人工神经网络、XGBoost这5种机器学习方法,基于美国华盛顿共享单车项目数据,分析时间因子、气象因子等对单车需求量的影响,实现对共享单车短时需求量的预测。仿真结果表明,影响单车需求量的主要因素包括温度、节假日、季节以及早晚高峰时间段等因素;极端随机树的预测效果最优,MAE和RMSE最小,为22.93和36.84,训练集得分和验证集得分最高,为1.0和0.941,与随机森林和其它算法相比,鲁棒性高,泛化能力强,且预测结果曲线与真实结果曲线相吻合,预测精度高,可为实际的车辆预测和调度提供参考依据。更多还原

【Abstract】 Shared bicycles are highly mobile and highly random, so it is important to quickly and accurately predict the short-term demand for shared bicycles in cities. Using five kinds of machine learning methods: random forest, extreme random tree, support vector machine, artificial neural network and XGBoost, and based on the shared bicycle project data in Washington, USA, we analyzed the influence of time factor and meteorological factors on the demand of bicycles, and predicted the sharing of bicycles. Short-term demand forecast. The simulation results show that the main factors affecting the demand for bicycle include temperature, holidays, seasons and morning and evening peak time periods; the extreme random tree has the best prediction effect, MAE and RMSE are the smallest, which are 22.93 and 36.84, training set score and verification set score are the highest, which are 1.0 and 0.941. Compared with random forests and other algorithms, the robustness is higher, the generalization ability is stronger, and the prediction result curve is consistent with the real result curve. The prediction accuracy is higher, which can be the actual vehicle prediction and scheduling.更多还原