We consider the problem of selecting covariates in a spatial regression model when the response is binary. Penalized likelihood-based approach is proved to be effective for both variable selection and estimation simultaneously. In the context of a spatially dependent binary variable, an uniquely interpretable likelihood is not available, rather a quasi-likelihood might be more suitable. We develop a penalized quasi-likelihood with spatial dependence for simultaneous variable selection and parameter estimation along with an efficient computational algorithm. The theoretical properties including asymptotic normality and consistency are studied under increasing domain asymptotics framework. An extensive simulation study is conducted to validate the methodology. Real data examples are provided for illustration and applicability. Although theoretical justification has not been made, we also investigate empirical performance of the proposed penalized quasi-likelihood approach for spatial count data to explore suitability of this method to a general exponential family of distributions.
The estimation of exceedance probabilities for extreme climatic events is critical for infrastructure design and risk assessment. Climatic events occur over a greater space than they are measured with point‐scale in situ gauges. In extreme value theory, the block maxima approach for spatial analysis of extremes depends on properly modeling the spatially varying Generalized Extreme Value marginal parameters (i.e., trend surfaces). Fitting these trend surfaces can be challenging since there are numerous spatial and temporal covariates that are potentially relevant for any given event type and region. Traditionally, covariate selection is based on assumptions regarding the topmost relevant drivers of the event. This work demonstrates the benefit of utilizing elastic‐net regression to support automatic selection from a relatively large set of physically relevant covariates during trend surface estimation. The trend surfaces presented are based on 24‐hr annual maximum precipitation for northeastern Colorado and the Texas‐Louisiana Gulf Coast.
more » « less- PAR ID:
- 10444522
- Publisher / Repository:
- DOI PREFIX: 10.1029
- Date Published:
- Journal Name:
- Geophysical Research Letters
- Volume:
- 49
- Issue:
- 11
- ISSN:
- 0094-8276
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Summary -
Extreme climatic events and human biology and health: A primer and opportunities for future researchmore » « less
Abstract Extreme climatic events are increasing in frequency, leading to hotter temperatures, flooding, droughts, severe storms, and rising oceans. This special issue brings together a collection of seven articles that describe the impacts of extreme climatic events on a diverse set of human biology and health outcomes. The first two articles cover extreme temperatures extending from extreme heat to cold and changes in winter weather and the respective implications for adverse health events, human environmental limits, well‐being, and human adaptability. Next, two articles cover the effects of exposures to extreme storms through an examination of hurricanes and cyclones on stress and birth outcomes. The following two articles describe the effects of extreme flooding events on livelihoods, nutrition, water and food insecurity, diarrheal and respiratory health, and stress. The last article examines the effects of drought on diet and food insecurity. Following a brief review of each extreme climatic event and articles covered in this special issue, I discuss future research opportunities–highlighting domains of climate change and specific research questions that are ripe for biological anthropologists to investigate. I close with a description of interdisciplinary methods to assess climate exposures and human biology outcomes to aid the investigation of the defining question of our time – how climate change will affect human biology and health. Ultimately, climate change is a water, food, and health problem. Human biologists offer a unique perspective for a combination of theoretical, methodological, and applied reasons and thus are in a prime position to contribute to this critical research agenda.
-
more » « less
Abstract Semi-competing risks refer to the time-to-event analysis setting, where the occurrence of a non-terminal event is subject to whether a terminal event has occurred, but not vice versa. Semi-competing risks arise in a broad range of clinical contexts, including studies of preeclampsia, a condition that may arise during pregnancy and for which delivery is a terminal event. Models that acknowledge semi-competing risks enable investigation of relationships between covariates and the joint timing of the outcomes, but methods for model selection and prediction of semi-competing risks in high dimensions are lacking. Moreover, in such settings researchers commonly analyze only a single or composite outcome, losing valuable information and limiting clinical utility—in the obstetric setting, this means ignoring valuable insight into timing of delivery after preeclampsia has onset. To address this gap, we propose a novel penalized estimation framework for frailty-based illness–death multi-state modeling of semi-competing risks. Our approach combines non-convex and structured fusion penalization, inducing global sparsity as well as parsimony across submodels. We perform estimation and model selection via a pathwise routine for non-convex optimization, and prove statistical error rate results in this setting. We present a simulation study investigating estimation error and model selection performance, and a comprehensive application of the method to joint risk modeling of preeclampsia and timing of delivery using pregnancy data from an electronic health record.
-
Like many other clinical and economic studies, each subject of our motivating transplant study is at risk of recurrent events of non-fatal tissue rejections as well as the terminating event of death due to total graft rejection. For such studies, our model and associated Bayesian analysis aim for some practical advantages over competing methods. Our semiparametric latent-class-based joint model has coherent interpretation of the covariate (including race and gender) effects on all functions and model quantities that are relevant for understanding the effects of covariates on future event trajectories. Our fully Bayesian method for estimation and prediction uses a complete specification of the prior process of the baseline functions. We also derive a practical and theoretically justifiable partial likelihood-based semiparametric Bayesian approach to deal with the analysis when there is a lack of prior information about baseline functions. Our model and method can accommodate fixed as well as time-varying covariates. Our Markov Chain Monte Carlo tools for both Bayesian methods are implementable via publicly available software. Our Bayesian analysis of transplant study and simulation study demonstrate practical advantages and improved performance of our approach.more » « less
-
more » « less
Summary Ecological momentary assessment is a method for collecting realtime data in subjects' environments. It often uses electronic devices to obtain information on psychological state through administration of questionnaires at times that are selected from a probability-based sampling design. This information can be used to model the effect of momentary variation in psychological state on the lifetimes to events such as smoking lapse. Motivated by this, a probability sampling framework is proposed for estimating the effect of time varying covariates on the lifetimes to events. Presented as an alternative to joint modelling of the covariate process as well as event lifetimes, this framework calls for sampling covariates at the event lifetimes and at times that are selected according to a probability-based sampling design. A design-unbiased estimator for the cumulative hazard is substituted into the log-likelihood, and the resulting objective function is maximized to obtain the proposed estimator. This estimator has two quantifiable sources of variation: that due to the survival model and that due to sampling the covariates. Data from a nicotine patch trial are used to illustrate the approach proposed.
