Research On Chinese Population Prediction Based On Leslie Model And Screening Of Influencing Factors

Population is the basic element of society and an important resource for regional development.The population problem has always been the core issue of the development of human society,and it is also a major challenge facing our country.With the development of social economy,the population data obtained by the national census reflecting Chinese characteristics has become more and more important.Since the national census work requires a lot of human,material and financial resources,it is a very meaningful task to predict population data.Through the established population prediction model,various data of the national census can be obtained more conveniently,which can save costs and improve efficiency.With the help of domestic and foreign population prediction theories and methods,using the population data of the "China Statistical Yearbook" released by the National Bureau of Statistics and referring to China’s population characteristics,this paper carries out research and exploration of Chinese prediction methods,aiming to effectively predict the Chinese population and obtain the data needed for the future national censuses through the research results.First,two new forecasting methods are constructed in Chinese population forecast.one is the forecasting method based on the Leslie model,and a Leslie model with data modeling is constructed.The model fits the fertility model with the log-normal distribution,predicts the sex ratio of the birth population with the gray GM(1,1)model,and improves the total fertility rate parameter according to the population policy,which is in line with the characteristics of the current population development plan.The other is to construct a regression model of influencing factors based on feature screening.Pearson correlation coefficient,gray correlation degree and XGBoost machine learning algorithm are used in the screening of influencing factors,and the main influencing factors are determined by combining three methods,and the modeling is carried out by linear regression analysis and nonlinear regression analysis.Empirical analysis shows that the fit of the two models is very ideal,and the model accuracy is high.Finally,the two models are used to predict the total population of China in the decade from 2021 to 2030.The results show that the population growth rate will be basically stable and the total national population will be expected to be 1.488 billion in 2030.Then the XGBoost machine learning algorithm is used to screen out the key influencing factors to construct a fitted model of influencing factors and time.And use the historical statistical data to train the BP neural network model,then use the fitting model to calculate the predicted values of the influencing factors of the male population and the urban population from 2021 to 2030,respectively,and input them into the trained BP neural network model,and then output the predicted values of the male population and urban population in China.The forecast results show that in 2030,China’s male population will be about 753 million,accounting for about 50.63%;in 2030,China’s urban population will be about 1.174 billion,accounting for about 78.90%.By combining the forecast results with the analysis of the official development planning report,it can be seen that the forecast results are in line with the actual development background and have high accuracy.Finally,the rationality of census frequency in China is investigated.Three data sampling frequencies of 5 years,10 years and 15 years are used,and regression analysis method was used to establish short-term,medium-term and long-term models of population prediction respectively.Comparing the results of the three models,the population prediction model with a sampling frequency of 10 years is the best,and this conclusion affirms the rationality of the current strategy of the national census with a frequency of 10 years from a scientific perspective.The innovations of this paper are as follows: it makes predictions on China’s population data in the new era according to the new situation,taking into account the latest population policies and population development plans.It builds an index system of influencing factors,and uses a combination of methods to comprehensively explore the impact of population Quantitative factors.From a statistical point of view,the rationality of the time and frequency of the national census is proved;the prediction method can be used to predict the total population,the proportion of various types of population and the proportion of urban population in the future in the short and medium term.The prediction results can be used as a great addition to national census data.