A study on the data fusion using the national health screening data

초록

The exact combination of data collected from different objects is difficult. Data fusion is the statistical combination process of obtaining an integrated dataset using common variable. We consider three statistical techniques for data fusion: conditional mean matching using linear regression models, gamma regression using nonlinear regression models on two independent datasets, and a distance hot deck nonparametric approach based on the distance of each variable. The National Health Insurance Corporation's National Health Screening Data are used to compare the performance of three models.

키워드