scispace - formally typeset
Search or ask a question

Showing papers by "Robert Tibshirani published in 1999"


Journal ArticleDOI
TL;DR: In CABG patients undergoing FTCA, the authors derived and validated risk factors of delayed extubation, prolonged ICU LOS, and mortality and developed a simplified CRS system with similar predictive abilities as the logistic models.
Abstract: BackgroundRisk factors of delayed extubation, prolonged intensive care unit (ICU) length of stay (LOS), and mortality have not been studied for patients administered fast-track cardiac anesthesia (FTCA). The authors’ goals were to determine risk factors of outcomes and cardiac risk scores (CRS) for

201 citations


Journal ArticleDOI
TL;DR: In this article, the covariance inflation criterion adjusts the training error by the average covariance of the predictions and responses, when the prediction rule is applied to permuted versions of the data set.
Abstract: We propose a new criterion for model selection in prediction problems. The covariance inflation criterion adjusts the training error by the average covariance of the predictions and responses, when the prediction rule is applied to permuted versions of the data set. This criterion can be applied to general prediction problems (e.g. regression or classification) and to general prediction rules (e.g. stepwise regression, tree-based models and neural nets). As a by-product we obtain a measure of the effective number of parameters used by an adaptive procedure. We relate the covariance inflation criterion to other model selection procedures and illustrate its use in some regression and classification problems. We also revisit the conditional bootstrap approach to model selection.

122 citations


Journal ArticleDOI
TL;DR: The findings relating to magnetic field exposures directly measured by personal monitoring support an association with the risk of childhood leukemia.
Abstract: Objectives: To evaluate the risk of childhood leukemia in relation to residential electric and magnetic field (EMF) exposures. Methods: A case control study based on 88 cases and 133 controls used different assessment methods to determine EMF exposure in the child's current residence. Cases comprised incident leukemias diagnosed at 0–14 years of age between 1985–1993 from a larger study in southern Ontario; population controls were individually matched to the cases by age and sex. Exposure was measured by a personal monitoring device worn by the child during usual activities at home, by point-in-time measurements in three rooms and according to wire code assigned to the child's residence. Results: An association between magnetic field exposures as measured with the personal monitor and increased risk of leukemia was observed. The risk was more pronounced for those children diagnosed at less than 6 years of age and those with acute lymphoblastic leukemia. Risk estimates associated with magnetic fields tended to increase after adjusting for power consumption and potential confounders with significant odds ratios (OR) (OR: 4.5, 95% confidence interval (CI): 1.3–15.9) observed for exposures 0.14 microTesla (μT). For the most part point-in-time measurements of magnetic fields were associated with non-significant elevations in risk which were generally compatible with previous research. Residential proximity to power lines having a high current configuration was not associated with increased risk of leukemia. Exposures to electric fields as measured by personal monitoring were associated with a decreased leukemia risk. Conclusions: The findings relating to magnetic field exposures directly measured by personal monitoring support an association with the risk of childhood leukemia. As exposure assessment is refined, the possible role of magnetic fields in the etiology of childhood leukemia becomes more evident.

111 citations


Journal ArticleDOI
TL;DR: A bootstrap-based method for enhancing a search through a space of models is proposed, well suited to complex, adaptively fitted models and provides a convenient method for finding better local minima and for resistant fitting.
Abstract: We propose a bootstrap-based method for enhancing a search through a space of models. The technique is well suited to complex, adaptively fitted models—it provides a convenient method for finding better local minima and for resistant fitting. Applications to regression, classification, and density estimation are described. We also provide results on the asymptotic behavior of bumping estimates.

63 citations


Journal ArticleDOI
02 Sep 1999-Nature
TL;DR: It is suggested that this illusion occurs because more time is generally spent being overtaken by other vehicles than is spent in overtaking them, and knowing that this effect is illusory might encourage drivers to resist small temptations to change lanes.
Abstract: The temptation to change lanes on a motorway may be prompted by an illusion. Switching lanes while driving along a busy road can be a risky manoeuvre. It is often instigated on the driver's judgement that the cars in the next lane are moving faster than those in the driver's own lane. But faulty intuition1,2,3 may cause people to overestimate the speed of vehicles in the next lane, believing that they are moving faster even when both lanes have the same average speed. We suggest that this illusion occurs because more time is generally spent being overtaken (passed) by other vehicles than is spent in overtaking them. Knowing that this effect is illusory might encourage drivers to resist small temptations to change lanes.

28 citations


Journal ArticleDOI
TL;DR: The proposed statistical measures for finding masses in mammograms are based on fitting broken line regressions to local intensity plots of the images and illustrate some of the statistical challenges in working with large diagnostic images.
Abstract: We propose statistical measures for finding masses in mammograms. The measures are based on fitting broken line regressions to local intensity plots of the images. The method is illustrated on a small database of mammograms that have been read by a radiologist and confirmed by operative data. This work illustrates some of the statistical challenges in working with large diagnostic images

8 citations