利用机器学习的国土空间城镇功能用地适宜性评价模型OA

To improve the suitability evaluation efficiency of urban function land in territorial space,we established a suitability evaluation index system for urban function land in territorial space.We used three machine learning algorithms,namely random forest,extreme gradient boosting,and convolutional neural network,to conduct suitability evaluation with Changsha City,Hunan Province,China as an example.The results show that the central area of Changsha City has high suitability,making it suitable for urban development.The western area has development potential with a larger number of suitable regions,while the eastern area has less suitable land due to harsh natural conditions.The random forest model shows lower prediction errors when handling feature interactions and nonlinear relationships,making it suitable for suitability evaluation in com-plex regions.Feature importance analysis indicates that reducing feature inputs had little effect on model accuracy,suggesting that feature selec-tion plays a crucial role in improving computational efficiency and reducing complexity.The research results can provide technical support for fu-ture territorial space planning,land resource management,and sustainable development.