An official website of the United States government
Summary This paper introduces an assumption-lean method that constructs valid and efficient lower predictive bounds for survival times with censored data. We build on recent work by Candès et al. (2023), whose approach first subsets the data to discard any data points with early censoring times and then uses a reweighting technique, namely, weighted conformal inference (Tibshirani et al., 2019), to correct for the distribution shift introduced by this subsetting procedure. For our new method, instead of constraining to a fixed threshold for the censoring time when subsetting the data, we allow for a covariate-dependent and data-adaptive subsetting step, which is better able to capture the heterogeneity of the censoring mechanism. As a result, our method can lead to lower predictive bounds that are less conservative and give more accurate information. We show that in the Type-I right-censoring setting, if either the censoring mechanism or the conditional quantile of the survival time is well estimated, our proposed procedure achieves nearly exact marginal coverage, where in the latter case we additionally have approximate conditional coverage. We evaluate the validity and efficiency of our proposed algorithm in numerical experiments, illustrating its advantage when compared with other competing methods. Finally, our method is applied to a real dataset to generate lower predictive bounds for users’ active times on a mobile app.
more »
« less
Accurately estimating correlations between demographic parameters: A comment on Deane et al. (2023)
Abstract Estimating correlations among demographic parameters is an important method in population ecology. A recent paper by Deane et al. (Ecology and Evolution13:e9847, 2023) attempted to explore the effects of different priors for covariance matrices on inference when using mark‐recovery data. Unfortunately, Deane et al. (2023) made a mistake when parameterizing some of their models. Rather than exploring the effects of different priors, they examined the effects of the use of incorrect equations on inference. In this manuscript, we clearly describe the mistake in Deane et al. (2023). We then demonstrate the use of an alternative and appropriate method and reach different conclusions regarding the effects of priors on inference. Consistent with other recent literature, informative inverse Wishart priors can lead to flawed inference, while vague priors on covariance matrix components have little impact when sample sizes are adequate.
more »
« less
- Award ID(s):
- 2209765
- PAR ID:
- 10609080
- Publisher / Repository:
- Wiley
- Date Published:
- Journal Name:
- Ecology and Evolution
- Volume:
- 14
- Issue:
- 9
- ISSN:
- 2045-7758
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
A recentRestoration Ecologyarticle by Merchant et al. (2022) suggested that practitioners do not regularly use functional traits in restoration planning. We disagree and provide our collective experience that practitioners do leverage trait‐based approaches and information, but in ways that are different from researchers. Here, we provide an expanded perspective that incorporates practitioner voices to provide a more complete assessment of how traits are used in restoration practice. We highlight that a major challenge in the field of restoration ecology that leads to a disconnect between researchers and practitioners is a different set of knowledge systems, goals, incentives, and limitations. We provide approaches that researchers can use to connect with practitioners and leverage their knowledge.more » « less
-
Since the first direct detection of gravitational waves by the LIGO–Virgo collaboration in 2015 (B. P. Abbott et al., 2016), the size of the gravitational-wave transient catalog has grown to nearly 100 events (R. Abbott et al., 2023), with the ongoing fourth observing run more than doubling the total number. Extracting astrophysical or cosmological information from these observations is a hierarchical Bayesian inference problem. GWPopulation is designed to provide simple-to-use, robust, and extensible tools for hierarchical inference in gravitational-wave astronomy or cosmology. It has been widely adopted for gravitational-wave astronomy, including producing flagship results for the LIGO-Virgo-KAGRA collaborations (Abac et al., 2024; R. Abbott et al., 2023). While designed to work with observations of compact binary coalescences, GWPopulation may be available to a wider range of hierarchical Bayesian inference problems.more » « less
-
Abstract We analyze spatial spreading in a population model with logistic growth and chemorepulsion. In a parameter range of short-range chemo-diffusion, we use geometric singular perturbation theory and functional-analytic farfield-core decompositions to identify spreading speeds with marginally stable front profiles. In particular, we identify a sharp boundary between between linearly determined, pulled propagation, and nonlinearly determined, pushed propagation, induced by the chemorepulsion. The results are motivated by recent work on singular limits in this regime using PDE methods (Grietteet al2023J. Funct. Anal.285110115).more » « less
-
Abstract Recent evidence suggests that community science and herbarium datasets yield similar estimates of species' phenological sensitivities to temperature. Despite this, two recent studies by Alecrim et al. (2023) and Miller et al. (2022) found very different results when using different data sources (community science and herbarium specimens, respectively) to investigate whether warming threatens wildflowers with phenological mismatch in relation to shading by deciduous trees.Here, we investigated whether differences between the two studies' results could be reconciled by testing four hypotheses related to model design, species, spatiotemporal data extent and phenophase.Hybrid model structures brought results from the two datasets closer together but did not fully reconcile the differences between the studies. Neither the species nor the phenophase selected for analysis seemed to be responsible for differences in results. Cropping the datasets to match spatial and temporal extents appeared to reconcile most differences but only at the cost of much higher uncertainty associated with reduced sample size.Synthesis: Our analysis suggests that although species‐level estimates of phenological sensitivity may be similar between community science and herbarium datasets, inherent differences in the types and extent of data may lead to contradictory inference about complex biotic interactions. We conclude that, until community science data repositories expand to match the range of climate conditions present in herbarium collections or until herbarium collections match the spatial extent and temporal frequency of community science repositories, ecological studies should ideally be evaluated using both datasets to test the possibility of biased results from either.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1002/ece3.70286
