Search

We investigate the problem of calibration and assessment of predictive rules in prognostic designs when missing values are present in the predictors. Our paper has two key objectives which are entwined. The first is to investigate how the calibration of the prediction rule can be combined with the use of multiple imputation to account for missing predictor observations. The second objective is to propose such methods that can be implemented with current multiple imputation software, while allowing for unbiased predictive assessment through validation on new observations for which outcome is not yet available. To inform the definition of methodology, we commence with a review of the theoretical background of multiple imputation as a model estimation approach as opposed to a purely algorithmic description. We specifically contrast application of multiple imputation for parameter (effect) estimation with predictive calibration. Based on this review, two approaches are formulated, of which the second utilizes application of the classical Rubin's rules for parameter estimation, while the first approach averages probabilities from models fitted on single imputations to directly approximate the predictive density for future observations. We present implementations using current software which allow for validatory or cross-validatory estimation of performance measures, as well as imputation of missing data in predictors on the future data where outcome is by definition as yet unobserved. We restrict discussion to binary outcome and logistic regression throughout, though the principles discussed are generally applicable. We present two data sets as examples from our regular consultative practice. Results show little difference between methods for accuracy but substantial reductions in variation of calibrated probabilities when using the first approach., Comment: This paper is under review at Biometrical Journal

We present the recent Fermilab calculations of the masses of the light quarks, using tadpole-improved Sheikholeslami-Wohlert (SW) quarks. Various sources of systematic errors are studied. Our final result for the average light quark mass in the quenched approximation evaluated in the $\bar{MS}$ scheme is $\bar{m}_q(\mu=2 GeV;n_f=0)= (m_u+m_d)/2=3.6 \pm 0.6 MeV$., Comment: 3 pgs. 3 figures. espcrc2.sty included. Talk presented at LATTICE96(phenomenology)

We report on the status of the determination of the heavy quark masses from our calculation of the quarkonia spectra. All sources of systematic errors that enter the quark mass determination are accounted for. We explicitly keep $ma \neq 0$ in the perturbative calculation relating the bare lattice mass to a renormalized mass. Our results are still preliminary., Comment: 3 pages, uuencoded postscript file, to appear in the proceedings of Lattice'94, Bielefeld, Germany, Sep. 27 - Oct. 1, 1994.

The renormalization of a general action for massive lattice fermions is discussed. The analysis applies for all $m_q a$. Preliminary results for the self energy at one loop in perturbation theory are presented., Comment: 3 pages (you guessed it: to appear in Lattice '93), uuencoded PostScript file, FERMILAB-CONF-93/382-T and EFI 93-72. Regular PostScript file by anonymous ftp to fnth06.fnal.gov as pub/Fermilab-Pub/93.382