Idle construction land prediction with Gradient Boosting Machine

Rapid urbanization has generated a large number of construction land problems in China, such as the idleness and illegal use of land. Whereas most methods have focused on the discovery of idle construction land by spatial overlay analysis, far less attention has been paid to the prediction of the idle construction land in advance. In this paper, a new method based on the Gradient Boosting Machine and relevant feature extraction is introduced to predict the status of construction land. Compared with other machine learning algorithms, Gradient Boosting Machine method achieves the best performance. It can be used to improve land use efficiency and avoid waste of land resources as well.