PDF(Pubmed)
Abstract：
A new dataset of population projections for local areas of Australia is described. The areas comprise SA3 areas of the Australian Statistical Geography Standard, which mostly range in population between 30,000 and 130,000. The projections are launched from the 2020 Estimated Resident Populations published by the Australian Bureau of Statistics and extend out to 2035. They are available by sex and five year age group up to 80-84 with 85+ as the final age group and in five year projection intervals. The projections were prepared using the synthetic migration cohort-component model, a new model for population projections which requires much less input data than conventional projection models, and therefore involves much lower costs and production time. Despite this, a recent evaluation demonstrated respectable forecast accuracy, and greater accuracy than equivalent simple projection models producing populations by age and sex. The age-sex projections are constrained to independent age-sex national projections and local area projections of total populations. The dataset consists of local area projections for the whole of the country which is consistent in methods, input data, and projection outputs due to the use of one model. This is rare in Australia because local area projections are most commonly prepared by individual State/Territory Governments using different methods, data sources, projection assumptions (which can be influenced by State/Territory population policies), and time periods. These nationally consistent projection data should be useful for a wide range of local area planning, policy, and research purposes, such as childcare demand, school enrolments, power and water usage, aged care provision, store and business site selection, living arrangements and household projections, labour force projections, and transport modelling.
摘要：
描述了澳大利亚当地人口预测的新数据集。这些区域包括澳大利亚统计地理标准的SA3区域，人口大多在30,000到130,000之间。这些预测是从澳大利亚统计局发布的2020年估计居民人口开始的，并延伸到2035年。它们按性别和5岁年龄组提供，最高可达80-84岁，最终年龄组为85岁，并以5年的预测间隔提供。预测是使用合成迁移队列-分量模型编制的，一种新的人口预测模型，比传统的预测模型需要少得多的输入数据，因此涉及更低的成本和生产时间。尽管如此,最近的一项评估显示了可观的预测准确性，比按年龄和性别产生人口的等效简单预测模型更准确。年龄-性别预测仅限于独立的年龄-性别国家预测和当地总人口预测。数据集包括全国的局部区域预测，这些预测在方法上是一致的，输入数据,和由于使用一个模型而产生的投影输出。这在澳大利亚很少见，因为地方预测通常是由各个州/地区政府使用不同的方法编制的，数据源,预测假设（可能受到州/地区人口政策的影响），和时间段。这些全国一致的预测数据应该对广泛的地方区域规划有用，政策,和研究目的，如儿童保育需求，学校入学率，电力和水的使用，老年护理提供，商店和商业网站的选择，生活安排和家庭预测，劳动力预测，和运输建模。
