基于LSTM神经网络的区域共享单车需求分析与预测

打开文本图片集

中图分类号：TP18 文献标识码：A文章编号：2096-4706（2025）22-0001-06

Regional Sharing Bicycle Demand Analysis and Prediction Based on LSTM

LUO Xiaoxuan， XIANG Zhongxue， XU Deheng，WANGBingkun，YU Pan （JiangxiInstituteofTechnology，Nanchang33oo98，China）

Abstract： Sharing bicycles make acertain contribution to alleviating traffic congestionand promoting green mobility，but the uneven alocation ofresources in actual operation can leadtothe problem ofdeclining servicequality.Acurate prediction of demads forsharing bicycles can optimize scheduling strategies and improve operational eficiency and service quality. Therefore，asharing bicycledemand prediction model isconstructed basedonLSTM.Thedata features are mainlyextracted byanalyzing theenvironmentalandtemporal factors，andfeatureengineringisperformed.ThenLSTMandRNNmodelsare constructedseparatelytocompare theirperformance.TheexperimentalresultsshowthattheLSTMmodeloutperforms theRNN modelin termsofdata fting and exhibitsstrong prediction performance.The modelcaneffectively assistoperators tooptimize vehicle scheduling，improve servicequalityanduser satisfaction，and provideareference basis foreficient managementof sharing bicycles.

Keywords： sharing bicycle;LSTM; future analysis; time-series prediction

0 引言

共享单车作为一种绿色、低碳的出行方式，在缓解城市交通拥堵、改善空气质量及解决“最后一公里”出行难题中具有重要作用[1]。（剩余7983字）