Application of imputation techniques in collaborative filtering-based recommender systems

This master thesis focuses on imputation methods as an approach to deal with the missing data problem. It's common that big datasets contain missing values. However, their existence may suppose a problem, as most of the usual statistical techniques can't be used and they may shadow importa...

Full description

Bibliographic Details
Author: Díaz Romo, Sandra
Format: master thesis
Publication Date:2018
Country:España
Institution:Universitat Oberta de Catalunya (UOC)
Repository:O2, repositorio institucional de la UOC
OAI Identifier:oai:openaccess.uoc.edu:10609/83565
Online Access:http://hdl.handle.net/10609/83565
Access Level:Open access
Keyword:imputation methods
recommender systems
multivariate analysis
mètode d'imputació
sistemes de recomanació
anàlisi multivariada
método de imputación
sistemas de recomendación
análisis multivariado
Computer algorithms -- TFM
Algorismes computacionals -- TFM
Algoritmos computacionales -- TFM
Description
Summary:This master thesis focuses on imputation methods as an approach to deal with the missing data problem. It's common that big datasets contain missing values. However, their existence may suppose a problem, as most of the usual statistical techniques can't be used and they may shadow important features, leading to the extraction of incorrect conclusions. In this context, imputation methods are statistical methods used to infer the missing values of a dataset using its intrinsic properties and the correlation amongst their variables. Several of these methods have been studied in detail and used in a recommender systems case study.