Open AccessResavanje problema nedostajucih podataka metodom visestrukog umetanja
Author(s) -
Vladimir Vasić
Publication year - 2003
Publication title -
economic annals/ekonomski anali
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.148
H-Index - 12
eISSN - 1820-7375
pISSN - 0013-3264
DOI - 10.2298/eka0357191v
Subject(s) - unavailability , missing data , sample (material) , statistics , mathematics , embedding , value (mathematics) , econometrics , computer science , physics , artificial intelligence , thermodynamics
In censuses as well as in various surveys, some respondents do not give answers to the stated questions. This, unavailability is a component of surveying. The problem arising from the sample unavailability is that the units selected to be included in the sample are, in fact, missing. Not only do these missing values cause less efficient estimates (due to the reduction in the sample size), but they also make impossible the use of standard statistical methods (which assume that there are no missing units). The multiple embedding method is a technique which replaces each missing or incomplete datum by two or more acceptable values drawn from the data distribution. It means that each lost value is replaced by m embedded values in the form of a vector.