Imputation through finite Gaussian mixture models

Zio, Marco Di; Guarnera, Ugo; Luzi, Orietta

doi:10.1016/j.csda.2006.10.002

Cited by 45 publications

(15 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Once the model that best fits data is selected and its parameters are estimated, for each incomplete observation y i = (y obs,i , y mis,i ), the conditional distribution f (y mis,i |y obs,i ; ) can be estimated as f y mis,i |y obs,i ;ˆ = K k=1τ ik N y mis,i |y obs,i ;θ k , and this probability distribution can be used for imputing missing values via its expected value (hereafter MCM) or through a random draw (MRD), as described in Di Zio et al (2007).…”

Section: Pmm Via Gaussian Mixtures Modelsmentioning

confidence: 99%

See 1 more Smart Citation

Semiparametric predictive mean matching

Zio

Guarnera

2008

AStA Adv Stat Anal

Self Cite

View full text Add to dashboard Cite

show abstract

Section: Pmm Via Gaussian Mixtures Modelsmentioning

confidence: 99%

“…The semiparametric predictive mean matching is compared to the nearest neighbor donor method and to model based imputations obtained via Gaussian mixture models as described in Di Zio et al (2007). The experiments are performed on both simulated and real data.…”

mentioning

confidence: 99%

Semiparametric predictive mean matching

Zio

Guarnera

2008

AStA Adv Stat Anal

Self Cite

View full text Add to dashboard Cite

show abstract

“…For example, standard complete-data methods can be applied directly, and the substantial effort required to create imputations needs to be carried out only once [7, 8]. On the other hand, multiple imputation generates a quantity of simulated values for each missing item, in order to reflect properly the uncertainty attached to missing data [9, 10]. This has been advocated as a statistically sound approach, but so far its use has been limited mainly to the social and medical sciences.…”

Section: Introductionmentioning

confidence: 99%

Clustering‐Based Multiple Imputation via Gray Relational Analysis for Missing Data and Its Application to Aerospace Field

Tian

et al. 2013

The Scientific World Journal

View full text Add to dashboard Cite

A large number of scientific researches and industrial applications commonly suffer from missing data. Some inappropriate techniques of missing value treatment compromise data quality, which detrimentally influences the knowledge discovery. In this paper, we propose a missing data completion method named CBGMI. Firstly, it separates the nonmissing data instances into several clusters by excluding the missing-valued entries. Then, it utilizes the entropy of the proximal category for each incomplete instance in terms of the similarity metric based on gray relational analysis. Experiments on UCI datasets and aerospace datasets demonstrate that the superiority of our algorithm to other approaches on validity.

show abstract

“…As shown in Fraley and Raftery [34,38], many probability distributions can be well approximated by mixture models. At the same time, in contrast to nonparametric schemes, mixture models do not require a large number of observations to obtain a good estimate [22,23].…”

Section: Imputation Via the Gmm Estimationmentioning

confidence: 99%

“…Priebe [22] shows that, with 10,000 observations, a log-normal density can be well approximated by a mixture of 30 Gaussian components. An empirical study by DiZio et al [23] shows that, for the preservation of the covariance structure, a random draw method is preferable over a conditional mean method when the GMM is used.…”

Section: Introductionmentioning

confidence: 99%

Imputation methods for dealing with missing scores in biometric fusion

Ding¹

View full text Add to dashboard Cite

Imputation Methods For Dealing with Missing Scores in Biometric Fusion Yaohui Ding I would like to express my sincere gratitude to Dr. Arun Ross for his technical and financial support throughout my study. The topic of this study is based on his idea. Without his wise suggestions and constant encouragement, I don't think I can enter and make any progress in this exciting field. I would also like to express my deep appreciation to the chair of the Department of Statistics, Dr. James Harner, for his guidance and help throughout my whole period of graduate education in Department of Statistics. It has been a great privilege to be his student. I would also like to acknowledge Dr. Mark Culp for serving in my thesis committee and providing me with valuable suggestions to improve this work. Several methods used in this thesis were based on his lectures in linear regression and categorical regression. I extend my special gratitude to all my colleagues, especially Brian, Aglika, Asem, Raghav and Raghu for their friendship and help to improve my work. Special thanks to my wife Jiajia. Without her understanding and care, this thesis would not have been possible.

show abstract

Imputation through finite Gaussian mixture models

Cited by 45 publications

References 8 publications

Semiparametric predictive mean matching

Semiparametric predictive mean matching

Clustering‐Based Multiple Imputation via Gray Relational Analysis for Missing Data and Its Application to Aerospace Field

Imputation methods for dealing with missing scores in biometric fusion

Contact Info

Product

Resources

About