基于改进的混沌鲸鱼算法优化DBN的广东人口数量预测研究  

Guangdong Population Prediction Based on Improved Chaotic Whale Algorithm for Optimizing DBN

作　　者：纪广月 JI Guang-yue(Guangdong University of Business and Technology,Zhaoqing 526020,China)

机构地区：[1]广东工商职业技术大学,广东肇庆526020

出　　处：《模糊系统与数学》2020年第2期164-174,共11页Fuzzy Systems and Mathematics

基　　金：广东省哲学社会科学“十三五”规划2017年度学科共建项目(GD17XSH03)。

摘　　要：为了估算广东省人口总数和老龄化人口总数,针对深度置信网络模型的性能易受其权值和阈值选择的影响,提出一种基于改进的混沌鲸鱼算法优化DBN的人口数量预测模型。首先,为提高鲸鱼算法的收敛速度和避免局部最优,将Skew Tent混沌模型和非线性收敛因子引入WOA算法用于初始化WOA种群和改进WOA更新策略。其次为了提高DBN模型的性能,运用ICWOA算法对DBN模型的权值和阈值进行优化选择,融合影响人口数量的多因素特征因子,将不同年龄段的总人数、死亡率、存活率、年生育率、第一产业占比、第二产业占比以及第三产业占比等多因素特征因子作为ICWOA-DBN的输入,人口数量作为ICWOA-DBN的输出,建立ICWOA-DBN人口数量预测模型。通过标准函数测试寻优对比发现,提出的ICWOA具有更快的收敛速度和更小的适应度值。为了验证本文算法ICWOA-DBN的预测性能,以第6次全国人口普查数据为参考依据,选择2005~2016年广东省历年常住人口总数和老龄化人口总数为研究对象,研究结果表明,与IWOA-DBN、WOA-DBN、GA-DBN和DBN相比,提出的ICWOA-DBN的人口数量预测模型的精度最高,为人口数量预测提供新的方法和途径。In order to estimate the total population of Guangdong Province and the total population of aging population, the performance of the deep confidence network model is susceptible to its weight and threshold selection. A population prediction model based on improved chaotic whale algorithm to optimize DBN is proposed. Firstly, in order to improve the convergence speed of whale algorithm and avoid local optimum, the Skew Tent chaotic model and nonlinear convergence factor are introduced into WOA algorithm to initialize WOA population and improve WOA update strategy. Secondly, in order to improve the performance of the DBN model, the ICWOA algorithm is used to optimize the weight and threshold of the DBN model, and the multi-factor characteristic factors affecting the population are integrated, and the total number of people, mortality, survival rate and annual fertility of different age groups will be combined. The multi-factor characteristic factors such as the rate, the proportion of the primary industry, the proportion of the secondary industry, and the proportion of the tertiary industry are input as ICWOA-DBN. The population is used as the output of ICWOA-DBN to establish the ICWOA-DBN population prediction model. Through the comparison of the standard function test optimization, the proposed ICWOA has a faster convergence speed and a smaller fitness value. In order to verify the prediction performance of the algorithm ICWOA-DBN, based on the 6 th national census data,the total number of permanent residents and the total number of aging population in Guangdong Province from 2005 to 2016 was selected as the research object. The research results show that with IWOA-DBN, WOA-DBN, GA-DBN and DBN, the proposed population prediction model of ICWOA-DBN has the highest accuracy and provides new methods and approaches for population prediction.

关 键 词：鲸鱼优化算法 深度置信网络 人口数量预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]