Existing Methods for Handling Incomplete Sample Clauses

Existing Methods for Handling Incomplete. Big Data Standard statistical methods for analyzing incomplete data, including the CC analysis, MI and IPW, gain their popularity on low dimensional data. However, in the presence of incomplete big data, their application is problematic especially when the data are complex and high-dimensional. In practice, it is very common to have large number of variables with missing values. In such cases, standard MI methods usually face severe chal- lenges because the imputation models for missing values are complicated and the intensive computing is most likely insurmountable. In the past two decades, with the increasing collection of data, modern techniques for handling incomplete big data have drawn great attention and both nonparametric and parametric methods have been developed.