Improved predictive mapping of indoor radon concentrations using ensemble regression trees based on automatic clustering of geological units.

PURPOSE: According to estimations around 230 people die as a result of radon exposure in Switzerland. This public health concern makes reliable indoor radon prediction and mapping methods necessary in order to improve risk communication to the public. The aim of this study was to develop an automated method to classify lithological units according to their radon characteristics and to develop mapping and predictive tools in order to improve local radon prediction. METHOD: About 240 000 indoor radon concentration (IRC) measurements in about 150 000 buildings were available for our analysis. The automated classification of lithological units was based on k-medoids clustering via pair-wise Kolmogorov distances between IRC distributions of lithological units. For IRC mapping and prediction we used random forests and Bayesian additive regression trees (BART). RESULTS: The automated classification groups lithological units well in terms of their IRC characteristics. Especially the IRC differences in metamorphic rocks like gneiss are well revealed by this method. The maps produced by random forests soundly represent the regional difference of IRCs in Switzerland and improve the spatial detail compared to existing approaches. We could explain 33% of the variations in IRC data with random forests. Additionally, the influence of a variable evaluated by random forests shows that building characteristics are less important predictors for IRCs than spatial/geological influences. BART could explain 29% of IRC variability and produced maps that indicate the prediction uncertainty. CONCLUSION: Ensemble regression trees are a powerful tool to model and understand the multidimensional influences on IRCs. Automatic clustering of lithological units complements this method by facilitating the interpretation of radon properties of rock types. This study provides an important element for radon risk communication. Future approaches should consider taking into account further variables like soil gas radon measurements as well as more detailed geological information.

On the robust measurement of inflectional diversity

Lexical diversity measures are notoriously sensitive to variations of sample size and recent approaches to this issue typically involve the computation of the average variety of lexical units in random subsamples of fixed size. This methodology has been further extended to measures of inflectional diversity such as the average number of wordforms per lexeme, also known as the mean size of paradigm (MSP) index. In this contribution we argue that, while random sampling can indeed be used to increase the robustness of inflectional diversity measures, using a fixed subsample size is only justified under the hypothesis that the corpora that we compare have the same degree of lexematic diversity. In the more general case where they may have differing degrees of lexematic diversity, a more sophisticated strategy can and should be adopted. A novel approach to the measurement of inflectional diversity is proposed, aiming to cope not only with variations of sample size, but also with variations of lexematic diversity. The robustness of this new method is empirically assessed and the results show that while there is still room for improvement, the proposed methodology considerably attenuates the impact of lexematic diversity discrepancies on the measurement of inflectional diversity.

Ultra-high Field (7 T) MRI and a Novel Robust Segmentation of Thalamic Nuclei on DWI for Gamma Knife Surgery Purposes: The Targeting of the Ventrointermediate Nucleus

Introduction: Gamma Knife surgery (GKS) is a noninvasive neurosurgical stereotactic procedure, increasingly used as an alternative to open functional procedures. This includes the targeting of the ventrointermediate nucleus of the thalamus (e.g., Vim) for tremor. Objective: To enhance anatomic imaging for Vim GKS using high-field (7 T) MRI and Diffusion Weighted Imaging (DWI). Methods: Five young healthy subjects and two patients were scanned both on 3 and 7 T MRI. The protocol was the same in all cases, and included: T1-weighted (T1w) and DWI at 3T; susceptibility weighted images (SWI) at 7T for the visualization of thalamic subparts. SWI was further integrated into the Gamma Plan Software® (LGP, Elekta Instruments, AB, Sweden) and co-registered with 3T images. A simulation of targeting of the Vim was done using the quadrilatere of Guyot. Furthermore, a correlation with the position of the found target on SWI and also on DWI (after clustering of the different thalamic nuclei) was performed. Results: For the 5 healthy subjects, there was a good correlation between the position of the Vim on SWI, DWI and the GKS targeting. For the patients, on the pretherapeutic acquisitions, SWI helped in positioning the target. For posttherapeutic sequences, SWI supposed position of the Vim matched the corresponding contrast enhancement seen at follow-up MRI. Additionally, on the patient's follow-up T1w images, we could observe a small area of contrast-enhancement corresponding to the target used in GKS (e.g., Vim), which belongs to the Ventral-Lateral-Ventral (VLV) nuclei group. Our clustering method resulted in seven thalamic groups. Conclusion: The use of SWI provided us with a superior resolution and an improved image contrast within the central gray matter, enabling us to directly visualize the Vim. We additionally propose a novel robust method for segmenting the thalamus in seven anatomical groups based on DWI. The localization of the GKS target on the follow-up T1w images, as well as the position of the Vim on 7 T, have been used as a gold standard for the validation of VLV cluster's emplacement. The contrast enhancement corresponding to the targeted area was always localized inside the expected cluster, providing strong evidence of the VLV segmentation accuracy. The anatomical correlation between the direct visualization on 7T and the current targeting methods on 3T (e.g., quadrilatere of Guyot, histological atlases, DWI) seems to show a very good anatomical matching.

Robust DNA Methylation in the Clonal Raider Ant Brain.

Social insects are promising model systems for epigenetics due to their immense morphological and behavioral plasticity. Reports that DNA methylation differs between the queen and worker castes in social insects [1-4] have implied a role for DNA methylation in regulating division of labor. To better understand the function of DNA methylation in social insects, we performed whole-genome bisulfite sequencing on brains of the clonal raider ant Cerapachys biroi, whose colonies alternate between reproductive (queen-like) and brood care (worker-like) phases [5]. Many cytosines were methylated in all replicates (on average 29.5% of the methylated cytosines in a given replicate), indicating that a large proportion of the C. biroi brain methylome is robust. Robust DNA methylation occurred preferentially in exonic CpGs of highly and stably expressed genes involved in core functions. Our analyses did not detect any differences in DNA methylation between the queen-like and worker-like phases, suggesting that DNA methylation is not associated with changes in reproduction and behavior in C. biroi. Finally, many cytosines were methylated in one sample only, due to either biological or experimental variation. By applying the statistical methods used in previous studies [1-4, 6] to our data, we show that such sample-specific DNA methylation may underlie the previous findings of queen- and worker-specific methylation. We argue that there is currently no evidence that genome-wide variation in DNA methylation is associated with the queen and worker castes in social insects, and we call for a more careful interpretation of the available data.

The environmental effects of changing speed limits: a quantile regression approach

Two speed management policies were implemented in the metropolitan area of Barcelona aimed at reducing air pollution concentration levels. In 2008, the maximum speed limit was reduced to 80 km/h and, in 2009, a variable speed system was introduced on some metropolitan motorways. This paper evaluates whether such policies have been successful in promoting cleaner air, not only in terms of mean pollutant levels but also during high and low pollution episodes. We use a quantile regression approach for fixed effect panel data. We find that the variable speed system improves air quality with regard to the two pollutants considered here, being most effective when nitrogen oxide levels are not too low and when particulate matter concentrations are below extremely high levels. However, reducing the maximum speed limit from 120/100 km/h to 80 km/h has no effect – or even a slightly increasing effect –on the two pollutants, depending on the pollution scenario. Length: 32 pages

Robust watermarking scheme based on the JPEG2000 standard

Peer reviewed

A Hierarchical Approach for Multi-task Logistic Regression

Peer-reviewed

Re-visiting the health care luxury good hypothesis: Aggregation, precision, and publication biases?

[cat] Mentre que una creixent literatura que ha examinat la relació entre la renda i la despesa sanitària suggereix que els serveis sanitaris són un be de luxe (elasticitat renda superior a la unitat), aquesta conclusió es contínuament debatuda atesa l'heterogeneïtat dels resultats. Aquest article testa la hipòtesis dels serveis sanitaris com bens de luxe fent server anàlisi de meta- regressió, particularment analitzant l'existència de biaixos de selecció de publicació, precisió així com biaixos d'agregació. Els resultats apunten l'existència d'un biaix de publicació, robust independentment dels controls analitzats. Els biaixos de precisió i agregació semblen tenir un paper en la generació de les estimacions de l'elasticitat renda. Els nostres resultat suggereixen que l'elasticitat renda dels serveis sanitaris un cop corregir pels biaixos esmentat varien entre 0.26 i 0.84, però no podem rebutjar que la elasticitat renda es igual a la unitat en algunes estimacions de l'elasticitat corregides.

Re-visiting the health care luxury good hypothesis: Aggregation, precision, and publication biases?

[cat] Mentre que una creixent literatura que ha examinat la relació entre la renda i la despesa sanitària suggereix que els serveis sanitaris són un be de luxe (elasticitat renda superior a la unitat), aquesta conclusió es contínuament debatuda atesa l'heterogeneïtat dels resultats. Aquest article testa la hipòtesis dels serveis sanitaris com bens de luxe fent server anàlisi de meta- regressió, particularment analitzant l'existència de biaixos de selecció de publicació, precisió així com biaixos d'agregació. Els resultats apunten l'existència d'un biaix de publicació, robust independentment dels controls analitzats. Els biaixos de precisió i agregació semblen tenir un paper en la generació de les estimacions de l'elasticitat renda. Els nostres resultat suggereixen que l'elasticitat renda dels serveis sanitaris un cop corregir pels biaixos esmentat varien entre 0.26 i 0.84, però no podem rebutjar que la elasticitat renda es igual a la unitat en algunes estimacions de l'elasticitat corregides.

Learning Preferences with Kernel-Based Methods

Learning of preference relations has recently received significant attention in machine learning community. It is closely related to the classification and regression analysis and can be reduced to these tasks. However, preference learning involves prediction of ordering of the data points rather than prediction of a single numerical value as in case of regression or a class label as in case of classification. Therefore, studying preference relations within a separate framework facilitates not only better theoretical understanding of the problem, but also motivates development of the efficient algorithms for the task. Preference learning has many applications in domains such as information retrieval, bioinformatics, natural language processing, etc. For example, algorithms that learn to rank are frequently used in search engines for ordering documents retrieved by the query. Preference learning methods have been also applied to collaborative filtering problems for predicting individual customer choices from the vast amount of user generated feedback. In this thesis we propose several algorithms for learning preference relations. These algorithms stem from well founded and robust class of regularized least-squares methods and have many attractive computational properties. In order to improve the performance of our methods, we introduce several non-linear kernel functions. Thus, contribution of this thesis is twofold: kernel functions for structured data that are used to take advantage of various non-vectorial data representations and the preference learning algorithms that are suitable for different tasks, namely efficient learning of preference relations, learning with large amount of training data, and semi-supervised preference learning. Proposed kernel-based algorithms and kernels are applied to the parse ranking task in natural language processing, document ranking in information retrieval, and remote homology detection in bioinformatics domain. Training of kernel-based ranking algorithms can be infeasible when the size of the training set is large. This problem is addressed by proposing a preference learning algorithm whose computation complexity scales linearly with the number of training data points. We also introduce sparse approximation of the algorithm that can be efficiently trained with large amount of data. For situations when small amount of labeled data but a large amount of unlabeled data is available, we propose a co-regularized preference learning algorithm. To conclude, the methods presented in this thesis address not only the problem of the efficient training of the algorithms but also fast regularization parameter selection, multiple output prediction, and cross-validation. Furthermore, proposed algorithms lead to notably better performance in many preference learning tasks considered.

Audit firm tenure and qualified opinions: New evidence from Spain

The 2010 Green Paper on Audit Policy by the European Commission has explicitly questioned the sufficiency of audit rotation rules established by European Union Members to guarantee auditor independence. In addition, the Paper clearly states that more research is needed regarding the effects of long audit tenures on independence. In this article, we have replicated the research by Ruiz-Barbadillo, Gómez-Aguilar, and Biedma (2005) about the effects of audit firm tenure on independence with more updated data. However, unlike them, we have performed panel data estimations instead of pooled regression. Our approach allows for a better control of individual unobserved heterogeneity, thus reducing potential problems caused by omitted variable bias. While Ruiz-Barbadillo et al. reported an unexpected positive effect of tenure on the likelihood of audit qualifications, we do not show any significant effect of tenure on the opinion of the audit report. Our results are robust to various sensitivity analyses.