摘 要:针对共享单车的站点投放量不平衡导致用户使用不便以及因共享单车使用率不高而使得企业运营困难的问题,提出一种基于 LightGBM 的共享单车短时需求量预测模型。首先研究天气、温度、时间等特征因素对共享单车短时使用量的影响,并通过提取主要特征有效降低模型的复杂度,然后采用贝叶斯优化对 LightGBM 进行调参建模,准确预测各站点每小时共享单车的需求量。最后通过实验将优化后的模型与基于 XGBoost 的模型和基于随机森林的模型进行对比,实验结果表明,优化后的模型更可靠,预测结果更精确。
关键词:共享单车;LightGBM;需求预测;特征分析;贝叶斯优化
DOI:10.19850/j.cnki.2096-4706.2022.20.021
基金项目:行为数据分析技术研究及应用(XTCXZX-2018-002)
中图分类号:TP18 文献标识码:A 文章编号:2096-4706(2022)20-0084-06
Short-term Demand Forecast of Shared Bicycle Based on LightGBM
LIU Benxing
(College of Computer Science and Technology, Taiyuan Normal University, Jinzhong 030619, China)
Abstract: Aiming at the problems of users’ inconvenience caused by the imbalance of the number of shared bicycles put on the site and the difficulty of enterprises’ operation caused by the low utilization rate of shared bicycles, a prediction model of shortterm demand for shared bicycles based on LightGBM is proposed. First, study the influence of weather, temperature, time and other characteristic factors on the short-term use of shared bicycles, and effectively reduce the complexity of the model by extracting the main characteristics. Then use Bayesian optimization to adjust the parameters of LightGBM, and accurately predict the hourly demand of shared bicycles at each station. Finally, the optimized model is compared with the XGBoost based model and the random forest based model through experiments. The experimental results show that the optimized model is more reliable and the prediction results are more accurate.
Keywords: shared bicycle; LightGBM; demand forecast; characteristic analysis; Bayesian optimization
参考文献:
[1] MA X L,TAO Z M,WANG Y H,et al. Long shortterm memory neural network for traffic speed prediction using remote microwave sensor data [J].Transportation Research Part C: Emerging Technologies,2015,54:187-197.
[2] XU C C,JI J Y,LIU P. The station-free sharing bike demand forecasting with a deep learning approach and large-scale datasets [J]. Transportation research part C: emerging technologies,2018,95:47-60.
[3] KALTENBRUNNER A,MEZA R,GRIVOLLA J,et al. Urban cycles and mobility patterns: Exploring and predicting trends in a bicycle-based public transport system [J].Pervasive and Mobile Computing,2010,6(4):455-466.
[4] 宋鹏,黄同愿,刘渝桥 . 基于 SVM 的共享单车需求预测[J]. 重庆理工大学学报(自然科学),2019,33(7):187-194.
[5] 曹旦旦,范书瑞,张艳,等 . 基于长短期记忆神经网络模型的共享单车短时需求量预测 [J]. 科学技术与工程,2020,20(20):8344-8349.
[6] 种颖珊,韩晓明 . 基于随机森林与时空聚类的共享单车站点需求量预测 [J]. 科学技术与工程,2018,18(32):89-94.
[7] 刘耿耿,朱予涵,郭灿阳 . 基于双向长短期记忆网络的共享单车流量预测 [J]. 小型微型计算机系统,2021,42(9):1871-1876.
[8] 冯易,王杜娟,胡知能,等 . 基于改进 LightGBM 集成模型的胃癌存活性预测方法 [J/OL]. 中国管理科学:1-15[2022-04-21].DOI:10.16381/j.cnki.issn1003-207x.2020.1628.
[9] 丁建立,孙玥.基于LightGBM的航班延误多分类预测 [J].南京航空航天大学学报,2021,53(6):847-854.
[10] 肖迁,焦志鹏,穆云飞,等 . 基于 LightGBM 的电动汽车行驶工况下电池剩余使用寿命预测 [J]. 电工技术学报,2021,36(24):5176-5185.
作者简介:刘本兴(1995—),女,汉族,河南信阳人,硕士研究生在读,研究方向:智能数据分析与应用。