Spatial prediction of population based on random forest

It is considered that there are difficulties in the study of population spatialization: the population data counted by the smallest administrative unit cannot intuitively show the characteristics of population distribution, and the speed and efficiency of traditional population spatialization models are low. In order to solve the above problems, this paper proposes to take Hubei Province as the research area, combine the two emerging geographic big data of points of interest and building outlines with traditional geographic data to extract features related to population settlements, and build and train a random forest model to predict population distribution at grid scale. The accuracy of the model performance is evaluated, and the results show that the simulation accuracy of this experiment is high, and the prediction accuracy is up to 0.902. It is proved that the processing speed and efficiency of the random forest model for multi-dimensional data is much higher than that of the traditional model, the discretization effect is better, and the purpose of fine-grained research can be achieved.