Variable selection for zero‐inflated and overdispersed data with application to health care demand in Germany

Wang, Zhu; Ma, Shuangge; Wang, Ching Yun

doi:10.1002/bimj.201400143

Cited by 44 publications

(55 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Consequently, we used either a negative binomial regression (Experiment 2) or a zero‐inflated negative binomial generalized linear model (ZINB; Experiment 3), while accounting for zero‐inflated data, to assess the number of trapped D. suzukii and pollinators by collection period, trap height, collection period × trap height, and month of collection . Because both models can account for overdispersion, we initially completed analyses using both models, and subsequently selected the model with the lowest Akaike information criterion (AIC) score . Interval duration (Experiment 2: 2, 3, 8 h; Experiment 3: 2, 3 7 h) data were used as the offset and the period with the highest trap catches was used as the reference variable.…”

Section: Methodssupporting

confidence: 65%

Vertical and temporal distribution of spotted‐wing drosophila (Drosophila suzukii) and pollinators within cultivated raspberries

Jaffe

Guédot

2019

Pest Management Science

View full text Add to dashboard Cite

BACKGROUND: Drosophila suzukii (Diptera: Drosophilidae) is the most prominent arthropod pest of caneberries. Current management practices rely on chemical control, which has raised concerns over the sustainability of this approach. We currently understand little about D. suzukii activity, and whether activity patterns can be exploited to improve management. In this study, we investigated the vertical and temporal distribution of D. suzukii, as well as pollinators, in cultivated raspberries. RESULTS: D. suzukii were generally crepuscular, and most active in the morning within the bottom half of the crop, and in the evening within the top half of the crop. Pollinators were most active during the day and within the top half of the canopy. Humidity and temperature were correlated with insect activity. CONCLUSION: Pesticides applied during twilight that ensure coverage in the bottom half of the crop, or if applied in the late afternoon cover the entire crop, should improve the pesticide efficacy.However, previous studies suggest that specific D. suzukii (e.g. sex, mating status, age) may exploit different areas in the crop at different times, and more work needs to be done to understand how these aspects before a specific management program can be recommended.

show abstract

Section: Methodssupporting

confidence: 65%

Vertical and temporal distribution of spotted‐wing drosophila (Drosophila suzukii) and pollinators within cultivated raspberries

Jaffe

Guédot

2019

Pest Management Science

View full text Add to dashboard Cite

show abstract

“…For the estimation of the model to predict future citations and to select variables we exclusively used the training cohort ( N = 259), whilst the test cohort ( N = 143) remained strictly reserved for model validation. All analyses were conducted using the statistical software r and extensions .…”

Section: Methodsmentioning

confidence: 99%

Predicting future citation counts of scientific manuscripts submitted for publication: a cohort study in transplantology

Kossmeier

Heinze

2018

Transpl Int

View full text Add to dashboard Cite

Citations are widely used for measuring scientific impact. The goal of the present study was to predict citation counts of manuscripts submitted to Transplant International (TI) in the two calendar years following publication. We considered a comprehensive set of 21 manuscript, author, and peer-review-related predictor variables available early in the peer-review process. We also evaluated how successfully the peer-review process at TI identified and accepted the most promising manuscripts for publication. A developed predictive model with nine selected variables showed acceptable test performance to identify often cited articles (AUROC = 0.685). Particularly important predictors were the number of pages, month of publication, publication type (review versus other), and study on humans (yes versus no). Accepted manuscripts at TI were cited more often than rejected but elsewhere published manuscripts (median 4 vs. 2 citations). The predictive model did not outperform the actual editorial decision. Both findings suggest that the peer-review process at TI, in its current form, was successful in selecting submitted manuscripts with a high scientific impact in the future. Predictive models might have the potential to support the review process when decisions are made under great uncertainty.

show abstract

“…Chun and Griffith (2013) list R code for stepwise selection in GLMs based on SAC minimization. In the mpath package (Wang et al 2015), the be.zeroinfl function performs a backward elimination (and forward selection) based on maximum likelihood criteria, and can be applied to zero-inflated models. Further variable selection algorithms for zero-inflated count data are presented in the medical literature, all proposing LASSO-based approaches (Chen et al 2016;Zeng et al 2014;Buu et al 2011) with different types of penalizations.…”

Section: A Backward Stepwise Algorithmmentioning

confidence: 99%

A Spatial-Filtering Zero-Inflated Approach to the Estimation of the Gravity Model of Trade

2018

View full text Add to dashboard Cite

Nonlinear estimation of the gravity model with Poisson-type regression methods has become popular for modelling international trade flows, because it permits a better accounting for zero flows and extreme values in the distribution tail. Nevertheless, as trade flows are not independent from each other due to spatial and network autocorrelation, these methods may lead to biased parameter estimates. To overcome this problem, eigenvector spatial filtering (ESF) variants of the Poisson/negative binomial specifications have been proposed in the literature on gravity modelling of trade. However, no specific treatment has been developed for cases in which many zero flows are present. This paper contributes to the literature in two ways. First, by employing a stepwise selection criterion for spatial filters that is based on robust (sandwich) p-values and does not require likelihood-based indicators. In this respect, we develop an ad hoc backward stepwise function in R. Second, using this function, we select a reduced set of spatial filters that properly accounts for importer-side and exporter-side specific spatial effects, as well as network effects, both at the count and the logit processes of zero-inflated methods. Applying this estimation strategy to a cross-section of bilateral trade flows between a set of 64 countries for the year 2000, we find that our specification outperforms the benchmark models in terms of model fitting, both considering the AIC and in predicting zero (and small) flows.

show abstract

Variable selection for zero‐inflated and overdispersed data with application to health care demand in Germany

Cited by 44 publications

References 23 publications

Vertical and temporal distribution of spotted‐wing drosophila (Drosophila suzukii) and pollinators within cultivated raspberries

Vertical and temporal distribution of spotted‐wing drosophila (Drosophila suzukii) and pollinators within cultivated raspberries

Predicting future citation counts of scientific manuscripts submitted for publication: a cohort study in transplantology

A Spatial-Filtering Zero-Inflated Approach to the Estimation of the Gravity Model of Trade

Contact Info

Product

Resources

About