ВЪЗМОЖНОСТИ ЗА ИЗПОЛЗВАНЕ НА МНОЖЕСТВЕНО ВЪВЕЖДАНЕ ПРИ ИГНОРИРУЕМИ МЕХАНИЗМИ
USAGE POSSIBILITIES OF MULTIPLE IMPUTATION UNDER IGNORABLE MECHANISMS

Summary/Abstract: The present research demonstrates the possibilities of multiple imputation as method for dealing with missing data. For this reason a data base is simulated, in which the variables form multivariate normal distribution. For the needs of the research in one of the variables artificially are implement missing values, which covers 50% of the values. Multiple imputation is conducted over data base with missing data and the derived results are in context of the “proper imputation” of Donald Rubin.

• Details
• Contents