Model-Based Clustering of Non-Gaussian Panel Data Based on Skew-<i>t</i>Distributions

Juárez, Miguel A.; Steel, Mark F. J.

doi:10.1198/jbes.2009.07145

Cited by 104 publications

(75 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Juarez & Steel (2010) find strong evidence in favor of heavier tails for a sample of OECD countries 1 . We investigate this economic growth illustration further and more formally in Section 6.…”

Section: Economic Growth Modelmentioning

confidence: 97%

“…The country-specific effect µ i can therefore be interpreted as the mean growth rate for country i. It is comparable across countries and corresponds to average covariate values; see Juarez & Steel (2010). This is a simple method for avoiding biases from correlation between x i,t and µ i .…”

Section: Student's T Dynamic Panel Data Modelmentioning

confidence: 99%

“…, T , where z i,t is the signal, ζ i,t is the Student's t distributed observation disturbance term, parameter γ measures state dependence, parameter vector β measures the effect of the explanatory variables x i,t whereas µ i captures unobserved countryspecific effects and ξ t captures the common time-varying effects. Variants of this model have been considered in for example Caselli et al (1996), Ho (2006) and Juarez & Steel (2010).…”

Section: Student's T Dynamic Panel Data Modelmentioning

confidence: 99%

See 2 more Smart Citations

Generalized dynamic panel data models with random effects for cross-section and time

Mesters¹,

Koopman²

2014

Journal of Econometrics

View full text Add to dashboard Cite

Section: Economic Growth Modelmentioning

confidence: 97%

Section: Student's T Dynamic Panel Data Modelmentioning

confidence: 99%

Section: Student's T Dynamic Panel Data Modelmentioning

confidence: 99%

See 1 more Smart Citation

Generalized dynamic panel data models with random effects for cross-section and time

Mesters¹,

Koopman²

2014

Journal of Econometrics

View full text Add to dashboard Cite

“…When using the t distribution, it is necessary to specify its scale parameter ν (read "nu"). We do not give that parameter a fixed value, but, again, vague priors, as suggested by Juárez and Steel (2010).…”

Section: Discussionmentioning

confidence: 99%

Brute force effects of mass media presence and social media activity on electoral outcome

Kovic¹,

Rauchfleisch²,

Metag³

et al. 2016

Preprint

View full text Add to dashboard Cite

In this study, we analyze whether the mere volume of presence in mass media and the mere volume of activity on social media convey advantages to candidates in parliamentary elections. Based on the theoretical model of bounded rationality, we call these potential effects brute force effects. During the last month of the election campaign of the Swiss federal election of 2015, we have tracked the presence of all 873 candidates in the canton of Zurich, the most populous canton, in a broad sample of mass media. Additionally, we have tracked those candidates' activity on Facebook and Twitter. The results of our multilevel Bayesian estimates show that mass media presence has a consistent non-trivial impact on different aspects of electoral outcome. Furthermore, social media activity also has a non-trivial impact, but only in terms of resonance (reactions to candidates' social media activity). Overall, our results suggest that brute force effects of of mass media presence and social media activity can have substantial impact on voting behavior.

show abstract

“…We complete the model statement by assigning the hyperpriors µ ν ∼ Beta(5, 5) and η ν ∼ Gamma(10, 0.1). Although these priors appear informative, when reparameterized, the prior specification is similar to the commonly used vague Gamma(2, 0.1) prior on the degrees of freedom ν t (Juárez and Steel, 2010). To regularize the robust data model, we modify the LASSO prior for the regression coefficients using the variance of the Student's t distribution, resulting in the prior…”

Section: Robust Regressionmentioning

confidence: 99%

Reconstruction of spatio-temporal temperature from sparse historical records using robust probabilistic principal component regression

Tipton

Hooten

Goring

2017

Adv. Stat. Clim. Meteorol. Oceanogr.

View full text Add to dashboard Cite

Abstract. Scientific records of temperature and precipitation have been kept for several hundred years, but for many areas, only a shorter record exists. To understand climate change, there is a need for rigorous statistical reconstructions of the paleoclimate using proxy data. Paleoclimate proxy data are often sparse, noisy, indirect measurements of the climate process of interest, making each proxy uniquely challenging to model statistically. We reconstruct spatially explicit temperature surfaces from sparse and noisy measurements recorded at historical United States military forts and other observer stations from 1820 to 1894. One common method for reconstructing the paleoclimate from proxy data is principal component regression (PCR). With PCR, one learns a statistical relationship between the paleoclimate proxy data and a set of climate observations that are used as patterns for potential reconstruction scenarios. We explore PCR in a Bayesian hierarchical framework, extending classical PCR in a variety of ways. First, we model the latent principal components probabilistically, accounting for measurement error in the observational data. Next, we extend our method to better accommodate outliers that occur in the proxy data. Finally, we explore alternatives to the truncation of lower-order principal components using different regularization techniques. One fundamental challenge in paleoclimate reconstruction efforts is the lack of out-of-sample data for predictive validation. Cross-validation is of potential value, but is computationally expensive and potentially sensitive to outliers in sparse data scenarios. To overcome the limitations that a lack of out-of-sample records presents, we test our methods using a simulation study, applying proper scoring rules including a computationally efficient approximation to leave-one-out cross-validation using the log score to validate model performance. The result of our analysis is a spatially explicit reconstruction of spatio-temporal temperature from a very sparse historical record.

show abstract

Model-Based Clustering of Non-Gaussian Panel Data Based on Skew-tDistributions

Cited by 104 publications

References 55 publications

Generalized dynamic panel data models with random effects for cross-section and time

Generalized dynamic panel data models with random effects for cross-section and time

Brute force effects of mass media presence and social media activity on electoral outcome

Reconstruction of spatio-temporal temperature from sparse historical records using robust probabilistic principal component regression

Contact Info

Product

Resources

About