ANALISIS REGRESI LINEAR GANDA DENGAN DATA HILANG MENGGUNAKAN IMPUTASI MULTIVARIAT CHAINED EQUATION (MICE); MULTIPLE LINEAR REGRESSION ANALYSIS WITH MISSING DATA USING MULTIVARIATE IMPUTATION CHAINED EQUATION(MICE)

PUSPITARINI, ERLINA, Danardono

2015 | Skripsi | FMIPA UGM

Missing data often occur in an observation. Often researchers discard observations containing missing data. But has found a variety of methods for dealing with missing data, among others, by imputation. Imputation consisted of imputing classical and modern imputation. Sometimes the value of imputation often unreasonable. Such amount is negative, especially if the case is categorical data will be very easy to very fatal error occurs. Therefore, it appears that a new method of imputation using chained equations (MICE), because the procedure is a chain so as to minimize the emergence of data that does not make sense. In the analysis performed iterations MICE after 5 times it will get the regression model fit as much as 5 times as well. Later than the fifth model fit inferred from giving a model called as a model pool.

Kata Kunci : missing data; imputation; chained equations; linear regression