A Study on The Population Projection of Taipei County with Time Series MOdels

碩士 === 國立臺北大學 === 統計學系 === 94 === The population is extremely important in whole environment system of one wreath, in index signs, such as many societies, economies and environment...etc., the fluctuation of the population usually can reflect the conditional interaction relation of the economic ener...

Full description

Bibliographic Details
Main Authors: Pei-Lin Chen, 陳珮玲
Other Authors: Shyang-Hua Wu
Format: Others
Language:zh-TW
Published: 2006
Online Access:http://ndltd.ncl.edu.tw/handle/50824155715163209333
Description
Summary:碩士 === 國立臺北大學 === 統計學系 === 94 === The population is extremely important in whole environment system of one wreath, in index signs, such as many societies, economies and environment...etc., the fluctuation of the population usually can reflect the conditional interaction relation of the economic energy and environment. The application level that study in light of the population gradual extension, the opportunity to use also increases, home currently relevant population estimate subject of research with the Council for Economic Development & Planning in Administration Yuan pushes to account of the population in the Taiwanese region number is lord, but it belongs to the population estimate of whole country, can't understand the individual condition of each county city, past origin the text aim be inquiring into the population of Taipei county and making the estimate to the short term. We research the population growth rate, crude birth rate, crude death rate and net migration rate and use the univariate ARIMA model, exponential smoothing model and transfer function model to construct the forecast model of population in Taipei county. The results find that the MAPE and THEIL of four models are not large, the forecast ability in-sample for the evaluate criterion RMSE and MAPE display that “the transfer function model of net migration rate for input variable” is better and the evaluate criterion THEIL display that “the transfer function model of crude birth rate for input variable” is better. And the forecast ability out-sample, the forecast ability for the evaluate criterion RMSE and MAPE display that “the transfer function model of net migration rate for input variable” is better and the evaluate criterion THEIL display that “the univariate ARIMA model” is better. With this some model estimates the populations of Taipei county, each mode all shows that the population in 2007, in January or February will have 3,740,000 people; Change to measure “the transfer function model of net migration rate for input variable” with the best list of the estimate ability to see, the population of Taipei county will have 3,767,000 people at the end of 2007, go to and will amount to 3,793,000 people at the end of 2008 .