A nonparametric asymptotic confidence interval for the median survival time is developed case where data are subject to arbitrary right censoring. This accomplished by inverting a generalization of sign test censored data. simulation study shows that this performs well variety underlying functions. The procedure applied from clinical trial compared f'our dosage regimens 5-uorouracil.

A method is developed for obtaining tree-structured relative risk estimates censored survival data. The first step of a full likelihood estimation procedure used in recursive partitioning algorithm that adopts most aspects the widely Classification and Regression Tree (CART) Breiman et al. (1984, Trees, Belmont, California: Wadsworth). performance technique investigated through stimulation compared to methods proposed by Davis Anderson (1989, Statistics Medicine 8, 947-961) Therneau,...

GRACE satellite gravity data is used to estimate terrestrial (surface plus ground) water storage within the Congo Basin in Africa for period of April, 2002–May, 2006. These estimates exhibit significant seasonal (30 ± 6 mm equivalent thickness) and long‐term trends, latter yielding a total loss ∼280 km 3 over 50‐month span data. We also combine precipitation sets (CMAP, TRMM) explore relative contributions source term hydrological balance Basin. find that tends saturate anomalies greater...

Glacial cycles redistribute water between oceans and continents causing pressure changes in the upper mantle, with consequences for melting of Earth's interior. Using Plio-Pleistocene sea-level variations as a forcing function, theoretical models mid-ocean ridge dynamics that include melt transport predict temporal crustal thickness hundreds meters. New bathymetry from Australian-Antarctic shows significant spectral energy near Milankovitch periods 23, 41, 100 ky, consistent model...

A new family of estimators a bivariate survival function based on censored vectors is obtained from decomposition the function. These are uniformly consistent under censoring and self-consistent univariate censoring.

Abstract In animal carcinogenesis studies, statistical methods that compare the production of occult tumors should tumor incidence rates. general, rate cannot be identified even approximately without both survival and sacrifice information. If information from frequent sacrifices is available, however, a function approximates can assumptions about relationship between death. Some earlier proposals based comparisons on prevalence rates other functions also compared under special conditions,...

A model of coupled mantle convection and planetary tectonics is used to demonstrate that history dependence can outweigh the effects a planet's energy content material parameters in determining its tectonic state. The convection-surface system allows multiple modes exist for equivalent parameter values. mode then determined by specific geologic climatic history. This implies models will not be able uniquely determine terrestrial planet without addition historical data. Historical data...

A topic that has received attention in both the statistical and medical literature is estimation of probability failure for endpoints are subject to competing risks. Despite this, it not uncommon see complement Kaplan–Meier estimate used this setting interpreted as failure. If one desires an can be way, however, cumulative incidence appropriate tool use such situations. We believe more commonly seen representations do lend themselves easy explanation understanding interpretation. present,...

Abstract Diagnostics for changes in the maximum likelihood estimate of β due to deletion single observations linear and logistic regression models are equivalent diagnostics obtained from fitting augmented models. For Cox proportional hazards model conditional arbitrarily matched or stratified data, derived again useful easy compute. These applicable different formulations multiplicative covariable effects (over time over strata), an exact approximate treatment tied failure times...

We present an analytic boundary layer model for thermal convection with a finite-strength plate and depth-dependent viscosity. The permits solutions in which convective flow rates the mantle exceed velocity. energy balance equations lithosphere cell are distinct yields velocity, thickness heat flow, as well laterally averaged horizontal profile cell. demonstrate, by inspection of lithospheric balance, that dominant driving force depends not only on material properties lithosphere, but also...

BACKGROUND A variety of innovative approaches to the prevention prostate cancer are now available, including selenium, alpha tocopherol, dietary interventions, and vitamin D. Perhaps most promising opportunity is based upon considerable evidence that cumulative androgen exposure contributes age-related risk cancer. METHODS The Prostate Cancer Prevention Trial has completed randomization over 18,000 healthy men either finasteride or placebo. CONCLUSIONS While primary objective this study...

Self-consistent estimators for estimating distribution functions from incomplete data are presented. In many cases these also generalized maximum likelihood estimators. this paper we discuss the theoretical properties of such estimators: existence, uniform consistency, law iterated logarithm, and weak convergence. Applications to product limit estimator right-censored proposed by Turnbull (1974, 1976) doubly (right- left-) censored given.

Self-consistent estimators for estimating distribution functions from incomplete data are presented. In many cases these also generalized maximum likelihood estimators. this paper we discuss the theoretical properties of such estimators: existence, uniform consistency, law iterated logarithm, and weak convergence. Applications to product limit estimator right-censored proposed by Turnbull (1974, 1976) doubly (right- left-) censored given.

Tarone and Ware (1977, Biometrika 64, 156-160) developed a general class of s-sample test statistics for right-censored survival data that includes the log-rank modified Wilcoxon procedures. Subsequently, many authors have considered two- classes in detail. In this paper family nonparametric is shown to unify existing generate new s (greater than or equal 2)-sample, trend, single continuous covariate problems.

A class of semi-Markov models, those which have proportional hazards and are forward-going (if state $j$ can be reached from $i$, then $i$ cannot $j$), shown to fit into the multiplicative intensity model counting processes after suitable random time changes. Standard large-sample results for following this therefore used make inferences on above including case where observations may censored. Large-sample a four-state in clinical trials presented.

Abstract A tree-based method for censored survival data is developed, based on maximizing the difference in between groups of patients represented by nodes a binary tree. The includes pruning algorithm with optimal properties analogous to classification and regression tree (CART) algorithm. Uniform convergence estimates conditional cumulative hazard functions discussed, an example given show utility developing prognostic classifications patients.