基于BP神经网络组合模型的云南省人口分类预测

发布时间：2018-01-22 17:10

本文关键词： 人口预测 GM(1 1)模型 Logistic模型 BP神经网络云南省少数民族　出处：《昆明理工大学》2016年硕士论文　论文类型：学位论文

【摘要】：人口问题一直是人类社会发展的关键问题,人口变化对于未来的经济、社会发展有着直接的重要影响。云南省是少数民族个数和总量较多的大省,解决好少数民族人口问题对于解决云南省的经济发展、社会稳定、民族团结等问题都大有好处。人口决策作为调控人口各项指标的直接手段,它的制定对经济社会发展起着举足轻重的作用,而人口准确预测直接关系到人口问题的科学决策,所以寻找一个更为科学、合理、准确的人口预测模型是历来学术界及政府制定决策的重要研究课题。进入21世界以来,云南省的总人口,少数民族人口形势都发生了很大的变化,并且各少数民族人口的数量规模参差不齐。面对新的社会环境,云南省人口及云南省少数民族人口还会如何发展,观点不一。本文试图通过GM(1,1)预测模型,Logistic预测模型以及BP神经网络组合预测模型对云南省总人口、云南省少数民族人口及极少数人口民族的人口进行中短期预测,研究云南省人口的主要特点,以期为云南省政府制定人口决策提供参考依据。GM(1,1)模型、Logistic模型、BP神经网络是常用的人口预测方法,本文首先介绍了GM(1,1)模型、Logistic模型、BP神经网络模型的理论部分,然后结合以往的理论和实践,考虑到组合预测模型的优势,文章基于GM(1,1)模型和Logistic模型两种单一预测模型建立了BP神经网络组合预测模型。然后选取1989-2014年的云南省人口数量变化特征,结合各个模型的特点选取合适的参数分别对云南省总人口、云南省少数民族人口、云南省极少数人口民族人口建立人口预测模型,进行数值模拟,分别对各个模型的预测结果的相对误差进行分析,对比分析每种模型的优劣,根据预测精度标准,选取预测精度最高的一种模型分别对云南省总人口、少数民族人口及极少数人口民族人口进行未来2015-2030年的人口预测。结果表明,BP神经网络组合预测模型预测精度最优,到2030年云南省人口总量、少数民族人口总量、极少数人口民族人口总量将分别为5005.03万人、1832.90万人、39.06万人。
[Abstract]:Population problem has always been the key issue of human social development. Population change has a direct and important impact on the future economic and social development. Yunnan Province is a large province with a large number and total number of ethnic minorities. To solve the problem of ethnic minority population is of great benefit to the economic development, social stability and national unity of Yunnan Province. Population decision-making as a direct means to control the population indicators. Its formulation plays an important role in the economic and social development, and accurate population prediction is directly related to the scientific decision-making of population issues, so looking for a more scientific and reasonable. Accurate population prediction model has always been an important research topic in academia and government decision-making. Since entering the 21st century, the situation of the total population and minority population in Yunnan Province has changed greatly. In the face of the new social environment, the population of Yunnan Province and the minority population of Yunnan Province will also develop, there are different views. This paper attempts to pass GM(1. 1) Logistic forecasting model and BP neural network combined forecasting model are used to predict the population of Yunnan Province in the short and medium term. The main characteristics of Yunnan population were studied in order to provide a reference basis for Yunnan provincial government to make population decision. BP neural network is a commonly used method for population prediction. This paper first introduces the theory of GM-1 / 1) Logistic model and BP neural network model. Then combined with the previous theory and practice, considering the advantages of the combined forecasting model, this paper based on GM(1. 1) BP neural network combined forecasting model is established by two single forecasting models, Logistic model and Logistic model. Then, the population change characteristics of Yunnan Province from 1989 to 2014 are selected. Combined with the characteristics of each model, appropriate parameters are selected to establish a population prediction model for the total population of Yunnan Province, the minority population of Yunnan Province, and the minority population of Yunnan Province, and the numerical simulation is carried out. The relative error of the prediction results of each model is analyzed, and the advantages and disadvantages of each model are compared. According to the prediction precision standard, the model with the highest prediction accuracy is selected to the total population of Yunnan Province. The population forecast for the future 2015-2030 is carried out by the minority population and the very minority national population. The results show that the BP neural network combined forecasting model has the best prediction accuracy. By 2030, the total population of Yunnan Province, the total population of ethnic minorities and the population of very minority nationalities will be fifty million fifty thousand and three hundred , eighteen million three hundred and twenty-nine thousand and 390,600 respectively.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：C924.2;TP183

【相似文献】