PYXIDA Institutional Repository
and Digital Library
Collections :

Title :Imputation methods based on principal component analysis
Alternative Title :Μέθοδοι που χρησιμοποιούνται για την αναπλήρωση ελλειπουσών τιμών και βασίζονται στην ανάλυση κύριων συνιστωσών
Creator :Siskas, Christos
Contributor :Papageorgiou, Ioulia (Επιβλέπων καθηγητής)
Athens University of Economics and Business, Department of Statistics (Degree granting institution)
Type :Text
Extent :80 p.
Language :en
Abstract :Principal Component Analysis is the oldest and most famous technique of Multivariate Analysis and can be used as a tool for researchers to deal with missingness in datasets. The aim of this thesis is the description, the analysis and the comparison of the techniques that belong in the category of Principal Component Analysis. All these available techniques are presented with respect to their theoretical framework and then a comparison of these methods in different percentages of missingness and for different types of datasets (simulated and real) follows in order to see which method responds better depending on the case and which is totally the most reliable.
Subject :Principal Components Analysis
Multivariate analysis
Fixed effect model
Random effect model
Date Issued :30-09-2016
Licence :

File: Siskas_2016.pdf

Type: application/pdf