Search PubMed⌕ Search
Home/Archive/PubMed

PubMed · 7786992

Influence diagnostics for generalized linear measurement error models.

Y Zhao·A H Lee·Y V Hui
View original source ↗Export metadata ↓

Abstract

We study influence diagnostics for generalized linear models when the true covariates are unobservable but measured with error. Based on the bias-corrected estimation of model parameters, diagnostic measures are developed to identify outlying and influential observations. The magnitude of influence is then assessed via a simulated envelope approach. The proposed diagnostic procedure is illustrated on two examples.

Explore related subjects

BiometryBlood GlucoseDiabetes MellitusHealth StatusHumansIncidenceLeast-Squares AnalysisMathematicsModels, StatisticalNorthern TerritoryProbabilityRecordsRegression AnalysisReproducibility of ResultsSportsComparative StudyJournal Article

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Y Zhao, A H Lee, Y V Hui. 1994. Influence diagnostics for generalized linear measurement error models.. https://pubmed.ncbi.nlm.nih.gov/7786992/

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed1997

The influence of the spatial structure of focal hepatic lesions on volume measurements performed with two and three-dimensional sonography using the ellipsoid formula.

A total of 63 patients with localized hepatic lesions were examined using two and three-dimensional sonography. The tumor volumes were calculated by means of the ellipsoid formula, compared with an experimentally determined reference volume and subsequently differentiated into correct and erroneous measurements according to the WHO criteria for tumor volumes. In a second test series, an analysis of the tumor shapes was carried out using parallel projections. It could be seen that 65% of the erroneous measurements resulted from the use of the ellipsoid formula, a procedure which is used to determine the volume of hepatic lesions. In the third part of this study, it was found that findings which are difficult to measure are frequently not recognized as such by the investigator making use of two-dimensional sonography. As a consequence, the volume measurements of focal hepatic lesions in oncology patients being considered for particular therapies should preferably be carried out with the aid of planimetric measurements rather than with the use of the ellipsoid formula.

Biometry↗