Eric Ruggieri

Change point method for detecting regime shifts in paleoclimatic time series: Application to δ18O time series of the Plio-Pleistocene

[1] Although different paleoenvironmental time series resolve past climatic change at different time scales, nearly all share one characteristic: they are nonstationary over the length of the record sampled. We describe a recursive dynamic programming change point algorithm that is well suited to identify shifts in the Earth system’s variability, as it represents a nonstationary time series as a series of regimes, each of which is homogeneous. The algorithm fits the data by minimizing squared errors not only over the parameters of the models for each subsequence but also over an arbitrary number of boundary points without restrictions on the lengths of regimes. The versatility of the algorithm is illustrated by an application to 5 Ma of Plio-Pleisotcene d 18 O variations. We seek to identify either the single dominant ‘‘Milankovitch’’ frequency or linear combinations of frequencies and consistently identify changes � 780 ka and � 2.7 Ma, among others, in each analysis done. Our applications also provide support to the recent hypothesis that obliquity-based Milankovitch terms can account for the circa 100 ka cycle that empirically dominates the most recent 1 million years.

An exact approach to Bayesian sequential change point detection

Change point models seek to fit a piecewise regression model with unknown breakpoints to a data set whose parameters are suspected to change through time. However, the exponential number of possible solutions to a multiple change point problem requires an efficient algorithm if long time series are to be analyzed. A sequential Bayesian change point algorithm is introduced that provides uncertainty bounds on both the number and location of change points. The algorithm is able to quickly update itself in linear time as each new data point is recorded and uses the exact posterior distribution to infer whether or not a change point has been observed. Simulation studies illustrate how the algorithm performs under various parameter settings, including detection speeds and error rates, and allow for comparison with several existing multiple change point algorithms. The algorithm is then used to analyze two real data sets, including global surface temperature anomalies over the last 130 years. Algorithm quickly updates its inference in linear time with each new observation.Derives uncertainty bounds on the number and location of change points in a data set.Explores potential detection criteria associated with posterior distribution.Simulation studies show high detection rate, low false positive rate.Analysis of two real data sets illustrate wide range of potential applications.

The Bayesian Change Point and Variable Selection Algorithm: Application to the δ 18 O Proxy Record of the Plio-Pleistocene

In this article, we introduce the Bayesian change point and variable selection algorithm that uses dynamic programming recursions to draw direct samples from a very high-dimensional space in a computationally efficient manner, and apply this algorithm to a geoscience problem that concerns the Earths history of glaciation. Strong evidence exists for at least two changes in the behavior of the Earths glaciers over the last five million years. Around 2.7 Ma, the extent of glacial cover on the Earth increased, but the frequency of glacial melting events remained constant at 41 kyr. A more dramatic change occurred around 1 Ma. For over three decades, the “Mid-Pleistocene Transition” has been described in the geoscience literature not only by a further increase in the magnitude of glacial cover, but also as the dividing point between the 41 kyr and the 100 kyr glacial worlds. Given such striking changes in the glacial record, it is clear that a model whose parameters can change through time is essential for the analysis of these data. The Bayesian change point algorithm provides a probabilistic solution to a data segmentation problem, while the exact Bayesian inference in regression procedure performs variable selection within each regime delineated by the change points. Together, they can model a time series in which the predictor variables as well as the parameters of the model are allowed to change with time. Our algorithm allows one to simultaneously perform variable selection and change point analysis in a computationally efficient manner. Supplementary materials including MATLAB code for the Bayesian change point and variable selection algorithm and the datasets described in this article are available online or by contacting the first author.

Effects of an In-House Coordinator and Practitioner Referral Rather Than Proxy Referral on Tissue Donation Rates

INTRODUCTION Timely referral of patients following asystolic death to an organ procurement organization (OPO) may increase tissue donation rates. Lack of education of health care providers and nonphysicians (admitting department) about timely referral to the OPO following asystolic death may adversely affect tissue donation rates. We hypothesized that using an in-house donation coordinator for provider education and changing the responsibility for calling the OPO from the admitting department to the licensed independent practitioner (LIP) declaring death would increase timely referral and tissue donation rates. METHODS An education program was developed in 2005 by a newly hired in-house coordinator to highlight the importance of tissue donation. In addition, to improve timely referrals to the OPO after death, the instructions accompanying the working copy of the death certificate were altered to require the patients LIP to call the OPO within 1 hour of death (early 2007). Rates for both timely referrals and tissue donors were modeled by a Poisson regression model with a log link function. RESULTS Timely referral rates rose from 48% before the interventions to 72% after the intervention (P < .0001). The number of tissue donors per number of referrals also increased significantly (P = .025) over that period. CONCLUSIONS An in-house donation coordinator initiated education program and LIP referral rather than referral by other parties following asystolic death results in higher tissue donation rates.

On efficient calculations for Bayesian variable selection

We describe an efficient, exact Bayesian algorithm applicable to both variable selection and model averaging problems. A fully Bayesian approach provides a more complete characterization of the posterior ensemble of possible sub-models, but presents a computational challenge as the number of candidate variables increases. While several approximation techniques have been developed to deal with problems that contain a large numbers of candidate variables, including BMA, IBMA, MCMC and Gibbs Sampling approaches, here we focus on improving the time complexity of exact inference using a recursive algorithm (Exact Bayesian Inference in Regression, or EBIR) that uses components of one sub-model to rapidly generate another and prove that its time complexity is O(m^2), where m is the number candidate variables. Testing against simulated data shows that EBIR significantly reduces compute time without sacrificing accuracy, while comparisons to the results obtained by MCMC approaches on the Crime and Punishment data set show that model averaging yields improved predictive performance over two model selection approaches. In addition, we show that finite mixtures of centroid solutions provide a means to better characterize the shape of multimodal posterior spaces than any individual model. Finally, we describe how the BIC approximations employed in the BMA and IBMA algorithms can be replaced by an EBIR calculation of equal time complexity and illustrate the departure of the BIC approximation from the exact Bayesian inference of EBIR.

A 10-Year Review of Total Hospital-Onset ICU Bloodstream Infections at an Academic Medical Center

Background The rates of central line‐associated bloodstream infections (CLABSIs) in U.S. ICUs have decreased significantly, and a parallel reduction in the rates of total hospital‐onset bacteremias in these units should also be expected. We report 10‐year trends for total hospital‐onset ICU‐associated bacteremias at a tertiary‐care academic medical center. Methods This was a retrospective analysis of all positive‐result blood cultures among patients admitted to seven adult ICUs for fiscal year 2005 (FY2005) through FY2014 according to Centers for Disease Control and Prevention National Healthcare Safety Network definitions. The rate of change for primary and secondary hospital‐onset BSIs was determined, as was the distribution of organisms responsible for these BSIs. Data from three medical, two general surgical, one combined neurosurgical/trauma, and one cardiac/cardiac surgery adult ICU were analyzed. Results Across all ICUs, the rates of primary BSIs progressively fell from 2.11/1,000 patient days in FY2005 to 0.32/1,000 patient days in FY2014; an 85.0% decrease (P < .0001). Secondary BSIs also progressively decreased from 3.56/1,000 to 0.66/1,000 patient days; an 81.4% decrease (P < .0001). The decrease in BSI rates remained significant after controlling for the number of blood cultures obtained and patient acuity. Conclusions An increased focus on reducing hospital‐onset infections at the academic medical center since 2005, including multimodal multidisciplinary efforts to prevent central line‐associated BSIs, pneumonia, Clostridium difficile disease, surgical site infections, and urinary tract infections, was associated with progressive and sustained decreases for both primary and secondary hospital‐onset BSIs.

Visualizing the Central Limit Theorem Through Simulation

Abstract The Central Limit Theorem is one of the most important concepts taught in an introductory statistics course, however, it may be the least understood by students. Sure, students can plug numbers into a formula and solve problems, but conceptually, do they really understand what the Central Limit Theorem is saying? This paper describes a simulation developed to help illustrate the Central Limit Theorem. Students use the computer mouse to hand draw a population of arbitrary shape and then watch as the sampling distribution grows with each sample selected. A simple assessment tool is also given to check students’ understanding of this crucial concept.