Construction method and device of population distribution prediction model, server and storage medium

Abstract

The invention provides a construction method of a population distribution prediction model, and belongs to the technical field of Internet. The method comprises the steps of optimizing a learning rate of each learning task and a parameter value of each model parameter based on a parameter value of each intermediate state in an optimization process of each model parameter, and obtaining an optimized learning rate of each learning task and an optimized parameter value of each model parameter; and adjusting the optimized parameter values of the model parameters based on the optimized learning rate of the target learning task to obtain a population distribution prediction model. According to the invention, meta-training is carried out based on the initial value of each model parameter and the initial learning rate of each learning task, the parameter value of the optimized model parameter containing meta-knowledge and the learning rate of the optimized learning task are obtained, and then the parameter value of the optimized model parameter is finely adjusted according to the learning rate of the optimized target learning task, so that the learning efficiency of the target learning task is improved. Therefore, a relatively accurate population distribution prediction model is constructed based on a small number of population prediction samples.

Description

technical field [0001] The present disclosure relates to the field of Internet technology, and in particular to a method, device, server and storage medium for constructing a population distribution prediction model. Background technique [0002] The ultra-fine-grained population estimation method refers to dividing the area that needs population estimation in a checkerboard pattern, counting the population of each grid point separately, and estimating the extremely small size through the statistical results of large-scale grid points (coarse-grained population distribution). (for example, 200m*200m or 100m*100m) grid point population number method. For target areas such as scenic spots, playgrounds, and parks with multiple points of interest, the long-term ultra-fine-grained population distribution prediction of the target area is conducive to optimizing the industrial structure of the target area and improving the economic benefits of the target area. [0003] In order to...

AI Technical Summary

Problems solved by technology

[0004] However, the cost of long-term ultra-fine-grained population distribution samples collected by GPS is high, and it is impossible to obtain a large n

Images