On the robustness of the generalized fused lasso to prior specifications

Using networks as prior knowledge to guide model selection is a way to reach structured sparsity. In particular, the fused lasso that was originally designed to penalize differences of coefficients corresponding to successive features has been generalized to handle features whose effects are structured according to a given network. As any prior information, the network provided in the penalty may contain misleading edges that connect coefficients whose difference is not zero, and the extent to which the performance of the method depend on the suitability of the graph has never been clearly assessed. In this work we investigate the theoretical and empirical properties of the adaptive generalized fused lasso in the context of generalized linear models. In the fixed \(p\) setting, we show that, asymptotically, adding misleading edges in the graph does not prevent the adaptive generalized fused lasso from enjoying asymptotic oracle properties, while forgetting suitable edges can be more problematic. These theoretical results are complemented by an extensive simulation study that assesses the robustness of the adaptive generalized fused lasso against misspecification of the network as well as its applicability when theoretical coefficients are not exactly equal. Our contribution is also to evaluate the applicability of the generalized fused lasso for the joint modeling of multiple sparse regression functions. Illustrations are provided on two real data examples.     

The West,Russia, and China in Central Asia: What Kind of Game Is Being Played in the Region?

The article discusses the nature of the tangled interaction between the ex-Soviet Central Asian republics and the outside great powers – Russia, the United States and China. Although the so-called Great Game model is often used to explain the geopolitical rivalry in the region, the current pattern of relations between various international actors is much more complex than a traditional tug-of-war between the old colonial empires. While Russia's strategic role in Central Asia appears to be on the rise, the capacity of the West to influence political developments in the region has diminished. However, the overall situation in Central Asia remains volatile and its geopolitical landscape is far from being settled.     

Bayesian unit-root tests for Stochastic Volatility models

In this article, we consider Bayesian inference procedures to test for a unit root in Stochastic Volatility (SV) models. Unit-root tests for the persistence parameter of the SV models, based on the Bayes Factor (BF), have been recently introduced in the literature. In contrast, we propose a flexible class of priors that is non-informative over the entire support of the persistence parameter (including the non-stationarity region). In addition, we show that our model fitting procedure is computationally efficient (using the software WinBUGS). Finally, we show that our proposed test procedures have good frequentist properties in terms of achieving high statistical power, while maintaining low total error rates. We illustrate the above features of our method by extensive simulation studies, followed by an application to a real data set on exchange rates.     

Determinanten des Informationsgehalts von Managementprognosen – eine empirische Untersuchung anhand der DAX und MDAX-Unternehmen

In this article the characteristics and determinants of the information content of management earnings forecasts of the German DAX and MDAX companies between 2002 and 2005 are analyzed. As proxy for the information content I use the precision of the wording of a forecast. I can show that the forecasts of MDAX companies are more precisely formulated than those of DAX companies. Furthermore forecasts are formulated more precisely, when companies have, (1) a lower percentage of intangible assets, (2) lower market capitalization, (3) lower volatility of earnings and (4) a higher need for external financing.     

Fathers’ Financial Support of Children in a Low Income Community in South Africa

We used data from the Birth to Twenty Cohort study to understand children’s receipt of financial support from their fathers in a low income, Black community in urban South Africa. Specifically, we (1) described fathers’ financial support over the life course of children; (2) estimated survival probabilities of receiving support for all children and not receiving support for children who experienced a parental union dissolution; and (3) identified factors that explained variation in the receipt of support after a union dissolution. Results suggest that most children received full or partial support throughout the life course. Furthermore, a high proportion of children received support after a union dissolution with much of the variation driven by pre-dissolution support, father’s education and the presence of extended kin.     

Urban land cover decreases the occurrence of a wetland endemic mammal and its associated vegetation

Urbanization and other land cover changes have been particularly detrimental to wetlands throughout the planet. One wetland specialist that may be sensitive to land cover changes surrounding wetlands is the round-tailed muskrat (Neofiber alleni; hereafter RTM). The RTM is a wetland obligate rodent that appears to have declined over the last half century and is a species of concern in Florida, where it is a near endemic. To determine if urbanization or other land cover influenced the distribution of RTMs we took a multi-scaled approach to examine the occurrence of RTMs and their associated vegetation in North-Central Florida. We detected RTMs on 19 of 72 sample plots and used a Classification And Regression Tree (CART) to determine that dogfennel (Eupatorium capillifolium) was negatively associated with RTMs and maidencane (Panicum hemitomon) was positively associated with their occurrence on sampling plots. Examining the influence of landscape composition for 2 km surrounding our plots we found that RTM occurrence was negatively related to urban land cover. Further, we found that dogfennel increased and maidencane decreased as urbanization increased in the surrounding landscape. Our research suggests that conservation of RTMs and their associated vegetation should focus on limiting urban sprawl at least within 2 km of wetlands.