基于因子与聚类分析的中国31省市的就业情况

Research on Chinese employment based on factor analysis and cluster analysis

数据深度作为一种多维的排序方法,已经被广泛应用于多元回归、聚类判别、风险度量等众多领域.基于《中国统计年鉴》(2010)的数据,在主成分分析的基础上,定义了新的深度函数作为评价准则对中国31省市的就业情况进行了综合排序.进一步以主成分的方差贡献率为权重,运用加权欧氏距离进行了主成分聚类.总结出4个层次的地区就业情况的总体特点与差异,并对其差异的形成原因进行了简要分析.

Data depths have been widely applied to many fields, such as clustering and classification, multivariate regression and risk measure. Based on ＂China statistical Yearbook＂ （2010） data and the principal component analysis, this paper ranks the employment of Chinese thirty-one provinces and cities with a new depth function. Regarding the variance contribution of the principal components as the weights, we classify the data with the principal component weighted clustering. It summarizes into four levels of employment with general characteristics and briefly analyzes the cause of the differences.