A multi-factor integrated school district school-age population forecasting method based on deep neural network

Abstract

The invention relates to a multi-factor fusion school age population forecasting method based on a depth neural network, which comprises the following steps: 1) preprocessing household registration data and provident fund data of each school district; 2) According to the preprocessed data, the distribution of population age, the length of school-age population moving in, the length of school-age population households moving in, the amount of provident fund contribution, and the time series of total population and school-age population are calculated, and the relative normalization treatment iscarried out; Using CNN and LSTM network to extract the temporal depth characteristics of total population and school-age population; 3) that total population time series depth characteristic, the school-age population time series depth characteristic, the population age distribution, the school-age population immigration time distribution, the school-age population household immigration time distribution, the provident fund contribution amount distribution and the like are spliced and sent into the whole connection layer to calculate the predicted value of the school-age population. The invention has important practical significance for reasonably planning educational investment and optimizing educational resource allocation.

Description

technical field [0001] The present invention relates to the field of population forecasting methods, in particular to a multi-factor fusion school-age population forecasting method based on deep neural networks. Background technique [0002] Population and education are two systems that have their own changes and are closely related. The school-age population has a profound impact on educational development (for example: educational needs, school layout, school buildings and teacher construction, etc.). However, the school-age population is affected by many factors, such as population migration, income, and regional education levels, etc., will all lead to changes in the school-age population, which poses great challenges to educational resource planning. To realize the rational allocation and orderly development of educational resources, it is necessary to grasp the law of the change of the school-age population and accurately predict the number of the school-age population...

AI Technical Summary

Problems solved by technology

This method of relying on domain knowledge modeling requires long-term design and verification, and will cause information loss to a certain extent, and the factors affecting the school-age population in different regions are quite different, making it difficult to adjust such models according to specific situations

Images