摘要
针对人口GIS中的海量人口数据,运用数据挖掘技术对人口信息的时空特征进行研究,挖掘人口数据背后的知识与规律。时间上,通过GM(1.1)、三次指数平滑与一元线性回归模型对人口信息变化趋势做定量研究,以近十年安徽全省年末户籍人口数量为例进行预测,发现安徽未来五年内人口总量平稳增长,人口压力较大。空间上,通过空间自相关理论对人口信息的聚集程度进行度量,以1997-2003年安徽区县人口增长率为例进行探测,依据Moran散点图,Moran’s I为0.1444,发现人口分布在整体上呈聚集状,具有较强的空间自相关。人口信息变化与经济发展、自然地貌有关,将人口信息诸多社会学特征一并纳入地域空间上考虑,通过数据挖掘能可以分析其时空演变规律。
In view of the massive population data in population GIS, conducts an application study on spatial and temporal characteristic of the population information with data mining technology to mine the knowledge and the rule behind population data. In terms of time, the demonstration shows that the year- end household quantity, population in Anhui province is about to be steady growth and the pressure will be huge in five years, with carrying on the population prediction by GM（1.1） model, three indices smooth as well as linear regress; In terms of space, then it explores the aggregation level of population information through the spatial autocorrelation theory, for example, taking a study of 1997- 2003 years county population growth, analyzing Moran scatter plot of population growth ratio of county city in Anhui province, the research obtains Moran ＇ s I is 0.1444, indicates the population distribute in the whole area forms the congregated shape,has the strong spatial autocorrelation . By the above study, population information change is related with economic development and natural landform. Integrating many sociology characteristics of the population information in the union region space, can be possible to analyze its space and time evolution rule through the data mining.
出处
《计算机技术与发展》
2009年第12期213-216,共4页
Computer Technology and Development
基金
国家自然科学基金项目(40771207)
安徽省高校自然科研项目(2006kj018B)
滁州学院校级科研项目(2008kj007B)
