地理空间信息2026，Vol.24Issue(6)：117-123,7.DOI:10.3969/j.issn.1672-4623.2026.06.021
利用LSTM-Transformer的城市小区入住率预测
Occupancy Rate Prediction of Urban Community Based on LSTM-Transformer:A Case Study of Jinan City,Shandong Province,China
摘要
Abstract
In the current stock era,the housing demand of urban residents tends to be diversified and personalized.The occupancy rate prediction of urban neighborhoods is not only important for real estate developers,governments and urban planners,but also can reveal the economic vitali-ty,population migration trend and social development potential of urban areas.In this paper,based on mobile phone signaling data combined with multi-scale socio-economic data,taking Jinan city,Shandong Province,China for example,we proposed an urban occupancy prediction model based on LSTM-Transformer,which was expected to fill in the gaps in the existing research,and explored the spatial distribution character-istics of occupancy rate in the city.The experimental result shows that the model is better than other traditional models in terms of prediction accuracy and has high reliability.The residential occupancy rate of Jinan City shows a decreasing trend from the central city to the outside,and the occupancy rates of the central and eastern areas are higher than that of the southwest.In terms of spatial distribution,there are high-value and low-value gathering areas in Jinan City,showing a change from"high-high"to"low-low"gathering mode from the downtown to the periphery.The prediction framework constructed in this study not only provides data support for precision real estate marketing in the stock era,but also pro-vides scientific basis and decision-making reference for the government to optimize urban public resource allocation,identify urban shrinkage risks,and promote fine-grained management of smart cities,which has significant social application value.关键词
入住率/AOI/LSTM-Transformer模型/时间序列预测Key words
occupancy rate/AOI/LSTM-Transformer model/time series prediction分类
天文与地球科学引用本文复制引用
王泽朝,周鸿运,吴政,马照亭,戴昭鑫..利用LSTM-Transformer的城市小区入住率预测[J].地理空间信息,2026,24(6):117-123,7.基金项目
国家重点研发计划资助项目(2024YFB2505804) （2024YFB2505804）
中央级公益性科研院所基本科研业务费资助项目(AR2504). （AR2504）
