统计研究

• 论文 • 上一篇    下一篇

特征样本重复抽样建模方法和应用研究

李宝瑜等   

  • 出版日期:2016-10-15 发布日期:2016-10-18

The Research on the Features Sample Resampling Regression Method and its Applications

Li Baoyu et al.   

  • Online:2016-10-15 Published:2016-10-18

摘要:

本文在传统统计回归方法的基础上,提出了一种新的FSR(Features Sample Resampling Regression)建模方法。该方法是依据变量特征采用机器抽样方法重复抽样,形成多个特征样本,对多个样本进行参数估计以后,形成参数的抽样分布,依据抽样分布,在多个优化目标要求下建立最优化模型。FSR方法能够作为一种社会科学研究中通用的建模方法应用。

关键词: 特征样本, 重复抽样, 小样本建模, 模型优化

Abstract:

On the basis of the traditional statistical regression method, this paper recommends a new modeling method called FSR (Features Sample Resampling Regression). FRS will implement machine sampling based on variable Features to form a lot of Features samples. After estimating the parameters of all samples, it gets the parameter distribution, based on which, the optimization model is established in optimization goals. The FSR method can be used as a general modeling method in social science research.

Key words: Features Samples, Repeated Sampling, Small Sample Modeling, Model Optimization