Recognizing Chinese Discourse Relations Based on Multi-Perspective and Hierarchical Modeling

Abstract Snow avalanches are natural hazards strongly controlled by the mountain winter climate, but their recent response to climate change has thus far been poorly documented. In this paper, hierarchical modeling is used to obtain robust indexes of the annual fluctuations of runout altitudes. The proposed model includes a possible level shift, and distinguishes common large-scale signals in both mean- and high-magnitude events from the interannual variability. Application to the data available in France over the last 61 winters shows that the mean runout altitude is not different now than it was 60 yr ago, but that snow avalanches have been retreating since 1977. This trend is of particular note for high-magnitude events, which have seen their probability rates halved, a crucial result in terms of hazard assessment. Avalanche control measures, observation errors, and model limitations are insufficient explanations for these trends. On the other hand, strong similarities in the pattern of behavior of the proposed runout indexes and several climate datasets are shown, as well as a consistent evolution of the preferred flow regime. The proposed runout indexes may therefore be usable as indicators of climate change at high altitudes.

Download Full-text

Practice Patterns for Older Adult Patients With Advanced Cancer: Physician Office Versus Hospital Outpatient Setting

Journal of Oncology Practice ◽

10.1200/jop.18.00315 ◽

2019 ◽

Vol 15 (1) ◽

pp. e30-e38 ◽

Cited By ~ 1

Author(s):

Allison Lipitz-Snyderman ◽

Coral L. Atoria ◽

Stephen M. Schleicher ◽

Peter B. Bach ◽

Katherine S. Panageas

Keyword(s):

Older Adult ◽

Advanced Cancer ◽

Odds Ratio ◽

Hierarchical Modeling ◽

Outpatient Setting ◽

Adult Patients ◽

Medicare Beneficiaries ◽

Hospital Outpatient ◽

Outpatient Departments ◽

Retrospective Comparative Study

PURPOSE: A shift in outpatient oncology care from the physician’s office to hospital outpatient settings has generated interest in the effect of practice setting on outcomes. Our objective was to examine whether medical oncologists’ prescribing of drugs and services for older adult patients with advanced cancer is used more in physicians’ offices compared with hospital outpatient departments. METHODS: This was a retrospective comparative study. SEER-Medicare data (2004 to 2011) were used to identify Medicare beneficiaries diagnosed with advanced breast, colon, esophagus, non–small-cell lung, pancreatic, or stomach cancer. Between physicians’ offices and hospital outpatient departments, we compared use of selected likely low-value supportive drugs, low-value therapeutic drugs, chemotherapy-related hospitalizations, and hospice. We used hierarchical modeling to assess differences between settings to account for correlation within physicians. RESULTS: Compared with patients treated in a hospital outpatient department, those treated in a physician’s office setting were more likely to receive erythropoiesis-stimulating agents (odds ratio, 1.72; 95% CI, 1.53 to 1.94) and granulocyte colony–stimulating factors (odds ratio, 1.28; 95% CI, 1.18 to 1.38). For combination chemotherapy and nanoparticle albumin-bound–paclitaxel in patients with breast cancer, there was a trend toward higher use in physicians’ offices, although this was not statistically significant. Chemotherapy-related hospitalizations and hospice did not vary by setting. CONCLUSION: We found somewhat higher use of several drugs for patients with advanced cancer in physicians’ office settings compared with hospital outpatient departments. Findings support research to dissect the mechanisms through which setting might influence physicians’ behavior.

Download Full-text

Apparent Survival Rates of Western Sandpiper (Calidris Mauri) Wintering in Northwest Baja California, Mexico

The Auk ◽

10.1093/auk/120.1.55 ◽

2003 ◽

Vol 120 (1) ◽

pp. 55-61 ◽

Cited By ~ 1

Author(s):

Guillermo Fernández ◽

Horacio de la Cueva ◽

Nils Warnock ◽

David B. Lank

Keyword(s):

Hierarchical Modeling ◽

Survival Rates ◽

Average Quality ◽

Apparent Survival ◽

Calidris Mauri ◽

Local Survival ◽

Western Sandpiper ◽

Western Sandpipers ◽

Best Fit ◽

Effect Of Age

AbstractTo estimate annual apparent local survival, we collected capture–resighting data on 256 individually marked male Western Sandpipers (Calidris mauri) wintering at Estero de Punta Banda, Mexico, between 1994–1997. A hierarchical modeling approach was used to address the effect of age class and year on survivorship rates. The best-fit model included a constant apparent survival probability (ϕ = 0.489; 95% CI = 0.410–0.569), but several models fit nearly as well, and averaging among the top five, to account for model uncertainty, suggested that adults had somewhat higher values than juveniles (ϕ = 0.490 ± 0.051 vs. 0.450 ± 0.067). Detection probability was substantially higher for adults than for juveniles (p = 0.741 vs. p = 0.537). Those apparent survival estimates are low compared with those from other studies of Western Sandpipers at breeding and other nonbreeding locations, and substantially lower than the true survivorship rates expected for small sandpipers in general. We interpret these results as indicating that this site is of below average quality for nonbreeding male Western Sandpipers.

Download Full-text

Variance prior specification for a basket trial design using Bayesian hierarchical modeling

Clinical Trials ◽

10.1177/1740774518812779 ◽

2018 ◽

Vol 16 (2) ◽

pp. 142-153 ◽

Cited By ~ 6

Author(s):

Kristen M Cunanan ◽

Alexia Iasonos ◽

Ronglai Shen ◽

Mithat Gönen

Keyword(s):

Standard Deviation ◽

Prior Distribution ◽

Hierarchical Modeling ◽

Bayesian Hierarchical Modeling ◽

Bayesian Hierarchical ◽

Inverse Gamma ◽

Wide Range ◽

Basket Trial ◽

Basket Trials ◽

Variance Parameter

Background: In the era of targeted therapies, clinical trials in oncology are rapidly evolving, wherein patients from multiple diseases are now enrolled and treated according to their genomic mutation(s). In such trials, known as basket trials, the different disease cohorts form the different baskets for inference. Several approaches have been proposed in the literature to efficiently use information from all baskets while simultaneously screening to find individual baskets where the drug works. Most proposed methods are developed in a Bayesian paradigm that requires specifying a prior distribution for a variance parameter, which controls the degree to which information is shared across baskets. Methods: A common approach used to capture the correlated binary endpoints across baskets is Bayesian hierarchical modeling. We evaluate a Bayesian adaptive design in the context of a non-randomized basket trial and investigate three popular prior specifications: an inverse-gamma prior on the basket-level variance, a uniform prior and half-t prior on the basket-level standard deviation. Results: From our simulation study, we can see that the inverse-gamma prior is highly sensitive to the input hyperparameters. When the prior mean value of the variance parameter is set to be near zero [Formula: see text], this can lead to unacceptably high false-positive rates [Formula: see text] in some scenarios. Thus, use of this prior requires a fully comprehensive sensitivity analysis before implementation. Alternatively, we see that a prior that places sufficient mass in the tail, such as the uniform or half-t prior, displays desirable and robust operating characteristics over a wide range of prior specifications, with the caveat that the upper bound of the uniform prior and the scale parameter of the half-t prior must be larger than 1. Conclusion: Based on the simulation results, we recommend that those involved in designing basket trials that implement hierarchical modeling avoid using a prior distribution that places a majority of the density mass near zero for the variance parameter. Priors with this property force the model to share information regardless of the true efficacy configuration of the baskets. Many commonly used inverse-gamma prior specifications have this undesirable property. We recommend to instead consider the more robust uniform prior or half-t prior on the standard deviation.

Download Full-text