Search

This paper describes the process of adapting the Stanford Coreference resolution module to the Basque language, taking into account the characteristics of the language. The module has been integrated in a linguistic analysis pipeline obtaining an end-to-end coreference resolution system for the Basque language. The adaptation process explained can benefit and facilitate other languages with similar characteristics in the implementation of their coreference resolution systems. During the experimentation phase, we have demonstrated that language-specific features have a noteworthy effect on coreference resolution, obtaining a gain in CoNLL score of 7.07 with respect to the baseline system. We have also analysed the effect that preprocessing has in coreference resolution, comparing the results obtained with automatic mentions versus gold mentions. When gold mentions are provided, the results increase 11.5 points in CoNLL score in comparison with results obtained when automatic mentions are used. The contribution of each sieve is analysed concluding that morphology is essential for agglutinative languages to obtain good performance in coreference resolution. Finally, an error analysis of the coreference resolution system is presented which have revealed our system’s weak points and help to determine the improvements of the system. As a result of the error analysis, we have enriched the Basque coreference resolution adding new two sieves, obtaining an improvement of 0.24 points in CoNLL F1 when automatic mentions are used and of 0.39 points when the gold mentions are provided. [ABSTRACT FROM AUTHOR]

BACKGROUND: The assessment of bone age and skeletal maturity and its comparison to chronological age is an important task in the medical environment for the diagnosis of pediatric endocrinology, orthodontics and orthopedic disorders, and legal environment in what concerns if an individual is a minor or not when there is a lack of documents. Being a time-consuming activity that can be prone to inter- and intra-rater variability, the use of methods which can automate it, like Machine Learning techniques, is of value. OBJECTIVE: The goal of this paper is to present the state of the art evidence, trends and gaps in the research related to bone age assessment studies that make use of Machine Learning techniques. METHOD: A systematic literature review was carried out, starting with the writing of the protocol, followed by searches on three databases: Pubmed, Scopus and Web of Science to identify the relevant evidence related to bone age assessment using Machine Learning techniques. One round of backward snowballing was performed to find additional studies. A quality assessment was performed on the selected studies to check for bias and low quality studies, which were removed. Data was extracted from the included studies to build summary tables. Lastly, a meta-analysis was performed on the performances of the selected studies. RESULTS: 26 studies constituted the final set of included studies. Most of them proposed automatic systems for bone age assessment and investigated methods for bone age assessment based on hand and wrist radiographs. The samples used in the studies were mostly comprehensive or bordered the age of 18, and the data origin was in most of cases from United States and West Europe. Few studies explored ethnic differences. CONCLUSIONS: There is a clear focus of the research on bone age assessment methods based on radiographs whilst other types of medical imaging without radiation exposure (e.g. magnetic resonance imaging) are not much explored in the literature. Also, socioeconomic and other aspects that could influence in bone age were not addressed in the literature. Finally, studies that make use of more than one region of interest for bone age assessment are scarce. [ABSTRACT FROM AUTHOR]

Autonomous Vehicles (AV) technology is emerging. Field tests on public roads have been on going in several states in the US as well as in Europe and Asia. During the US public road tests, crashes with AV involved happened, which becomes a concern to the public. Most previous studies on AV safety relied heavily on assessing drivers’ performance and behaviors in a simulation environment and developing automated driving system performance in a closed field environment. However, contributing factors and the mechanism of AV-related crashes have not been comprehensively and quantitatively investigated due to the lack of field AV crash data. By harnessing California’s Report of Traffic Collision Involving an Autonomous Vehicle Database, which includes the AV crash data from 2014 to 2018, this paper investigates by far the most current and complete AV crash database in the US using statistical modeling approaches that involve both ordinal logistic regression and CART classification tree. The quantitative analysis based on ordinal logistic regression and CART models has successfully explored the mechanism of AV-related crash, via both perspectives of crash severity and collision types. Particularly, the CART model reveals and visualize the hierarchical structure of the AV crash mechanism with knowledge of how these traffic, roadway, and environmental contributing factors can lead to crashes of various serveries and collision types. Statistical analysis results indicate that crash severity significantly increases if the AV is responsible for the crash. The highway is identified as the location where severe injuries are likely to happen. AV collision types are affected by whether the vehicle is on automated driving mode, whether the crashes involve pedestrians/cyclists, as well as the roadway environment. The method used in this research provides a proven approach to statistically analyze and understand AV safety issues. And this benefit is potential be even enhanced with an increasing sample size of AV-related crashes records in the future. The comprehensive knowledge obtained ultimately facilitates assessing and improving safety performance of automated vehicles. [ABSTRACT FROM AUTHOR]

European farmlands are vital arenas for socio-ecological sustainability because of their significant land coverage and potential for integrating food production with biodiversity conservation. The knowledge produced by scientific research is a critical ingredient in developing and implementing socio-economically and ecologically sustainable management strategies for farming landscapes. The grasslands of Europe have been managed for millennia. They have exceptional socio-cultural and economic value and are among the most biodiverse ecosystems in the world. The quality of scientific knowledge on them and its potential to address grasslands as complex socio-ecological systems is strongly dependent not only on the creativity and scientific ambition of the researcher, but also on the network around the researcher (including both academic and non-academic sectors). The goal of this study is to map the research network around Romania’s grasslands using bibliometrics analysis, a well-developed scientific domain that utilizes network theory to analyze relationships between affiliations networks, co-authorship networks, and co-word analysis. The number of studies targeting grasslands in Romania is increasing, owing mostly to international involvement. However, management of the grasslands is still deficient and the contribution of science to the process is virtually absent. The current research is mainly related to the biological and ecological characteristics of grasslands, with topics related to their management notably absent from internationally visible research, especially in the context of EU Common Agricultural Policies. To increase scientific inquiry and better inform the EU and local policies on grasslands management, Romanian researchers should capitalize on international collaborations and local academic leaders. Our findings can be used to identify research gaps and to improve collaboration and knowledge exchange between practitioners, researchers, policy makers, and stakeholders. [ABSTRACT FROM AUTHOR]

Commercially motivated junk news–i.e. money-driven, highly shareable clickbait with low journalistic production standards–constitutes a vast and largely unexplored news media ecosystem. Using publicly available Facebook data, we compared the reach of junk news on Facebook pages in the Netherlands to the reach of Dutch mainstream news on Facebook. During the period 2013–2017 the total number of user interactions with junk news significantly exceeded that with mainstream news. Over 5 Million of the 10 Million Dutch Facebook users have interacted with a junk news post at least once. Junk news Facebook pages also had a significantly stronger increase in the number of user interactions over time than mainstream news. Since the beginning of 2016 the average number of user interactions per junk news post has consistently exceeded the average number of user interactions per mainstream news post. [ABSTRACT FROM AUTHOR]

Recent years have seen tremendous advances in the scientific study of networks, as more and larger data sets of relationships among nodes have become available in many different fields. This has led to pathbreaking discoveries of near-universal network behavior over time, including the principle of preferential attachment and the emergence of scaling in complex networks. Missing from the set of network analysis methods to date is a measure that describes for each node how its relationship (or links) with other nodes changes from one period to the next. Conventional measures of network change for the most part show how the degrees of a node change; these are scalar comparisons. Our contribution is to use, for the first time, the cosine similarity to capture not just the change in degrees of a node but its relationship to other nodes. These are vector (or matrix)-based comparisons, rather than scalar, and we refer to them as “rewiring” coefficients. We apply this measure to three different networks over time to show the differences in the two types of measures. In general, bigger increases in our rewiring measure are associated with larger increases in network density, but this is not always the case. [ABSTRACT FROM AUTHOR]

Objectives: In August 2016, new prescription guidelines were introduced in Hungary to reduce the co-payments for antibiotics among children aged 0-4. This study aims at analysing the implementation of this policy and its effect on the use of antibiotics. Methods: The analysis is based on administrative prescription records between January 2010—February 2018, covering the entire population of Hungary aged 0-7. Spatial autocorrelation indices are calculated and settlement level regression models are estimated to analyse the spatial variation in the application of the new guidelines. The effect of reduced co-payments on antibiotic use is estimated with a difference-in-differences type model: the treatment and control groups are children aged 0-4 and 5-7, respectively; the treatment and control periods are August 2016—February 2018 and January 2010—July 2016, respectively. Results: The new prescription guidelines are more widely applied in settlements with higher per capita income and lower unemployment rate. Adherence to the new guidelines is spatially clustered. A 10–15% decrease in the out-of-pocket costs of antibiotics is estimated to increase the consumption of antibiotics by about 5% (95% CI: 2.63%–7.55%). Conclusions: In the absence of clear enforcement mechanisms, the adoption of the new prescription guidelines is selective, contradicting the aims of the policy of making antibiotics affordable for the poor children. The results point to the possible role of physicians’ information networks in the application of prescription guidelines. The use of antibiotics among children aged 0-4 is responsive to the price subsidy of antibiotics. [ABSTRACT FROM AUTHOR]

Sustainable development goals are used as a guidance for strategies development on local, regional and national levels. The importance of including young people in this complex process is recognized in all relevant documents (i.e. Agenda 21), however it is not an easy task to elicit opinions and preferences from the youth. Furthermore, the assessment of the sustainable development goals itself presents a challenge for the noisy data and nonlinear relationships in data. Popular approach is fuzzy set models where expert knowledge is presented with comprehensible rules; however expert knowledge elicitation takes a long time too. Several studies proposed an adaptive neuro-fuzzy inference system approach that combines the fuzzy set theory to model expert knowledge with neural networks for inferring rules and membership functions from data to assess the sustainable development performance. We base our assumptions that ANFIS can be used to predict the importance of sustainable development pillars from the demographic data of young people. For this purpose, we have conducted an online survey on sustainable development goals opinions and importance of young people in Serbia. The sample of 386 respondents has been split into a training sample of 300 instances (to generate membership functions and fuzzy rules) and a testing sample of 86 instances to predict the importance of the three pillars. We have conducted a trace-driven simulation test to validate the results of the proposed ANFIS model. Results of the study provided insights into how the young people in Serbia assess the importance of sustainable development goals. Secondly, the results suggest that ANFIS can be applied to predict values of importance of the three sustainable development pillars with the relative error of Rel Err < 5%. It must be noted that the considered model could be further improved by using training samples with more data. [ABSTRACT FROM AUTHOR]

The topology of animal transport networks contributes substantially to how fast and to what extent a disease can transmit between animal holdings. Therefore, public authorities in many countries mandate livestock holdings to report all movements of animals. However, the reported data often does not contain information about the exact sequence of transports, making it impossible to assess the effect of truck sharing and truck contamination on disease transmission. The aim of this study was to analyze the topology of the Swiss pig transport network by means of social network analysis and to assess the implications for disease transmission between animal holdings. In particular, we studied how additional information about transport sequences changes the topology of the contact network. The study is based on the official animal movement database in Switzerland and a sample of transport data from one transport company. The results show that the Swiss pig transport network is highly fragmented, which mitigates the risk of a large-scale disease outbreak. By considering the time sequence of transports, we found that even in the worst case, only 0.34% of all farm-pairs were connected within one month. However, both network connectivity and individual connectedness of farms increased if truck sharing and especially truck contamination were considered. Therefore, the extent to which a disease may be transmitted between animal holdings may be underestimated if we only consider data from the official animal movement database. Our results highlight the need for a comprehensive analysis of contacts between farms that includes indirect contacts due to truck sharing and contamination. As the nature of animal transport networks is inherently temporal, we strongly suggest the use of temporal network measures in order to evaluate individual and overall risk of disease transmission through animal transportation. [ABSTRACT FROM AUTHOR]

Objective: To explore pregnant women’s preferences for birth setting in England. Design: Labelled discrete choice experiment (DCE). Setting: Online survey. Sample: Pregnant women recruited through social media and an online panel. Methods: We developed a DCE to assess women’s preferences for four hypothetical birth settings based on seven attributes: reputation, continuity of care, distance from home, time to see a doctor, partner able to stay overnight, chance of straightforward birth and safety for baby. We used a mixed logit model, with setting modelled as an alternative-specific constant, and conducted a scenario analysis to evaluate the impact of changes in attribute levels on uptake of birth settings. Main outcome measures: Women’s preferences for birth setting. Results: 257 pregnant women completed the DCE. All birth setting attributes, except ‘time to see doctor’, were significant in women’s choice (p<0.05). There was significant heterogeneity in preferences for some attributes. Changes to levels for ‘safety for the baby’ and ‘partner able to stay overnight’ were associated with larger changes from baseline uptake of birth setting. If the preferences identified were translated into the real-world context up to a third of those who reported planning birth in an obstetric unit might choose a midwifery unit assuming universal access to all settings, and knowledge of the differences between settings. Conclusions: We found that ‘safety for the baby’, ‘chance of a straightforward birth’ and ‘can the woman’s partner stay overnight following birth’ were particularly important in women’s preferences for hypothetical birth setting. If all birth settings were available to women and they were aware of the differences between them, it is likely that more low risk women who currently plan birth in OUs might choose a midwifery unit. [ABSTRACT FROM AUTHOR]

Background: The Hopkins Symptom Checklist in 25 items (HSCL-25) helps to assess anxiety and depression in Primary Care. Anxiety and depression show considerable overlap in primary care. This self-administrated questionnaire is valid, reliable and ergonomic in the original US version. We have translated it into French. The aim of this study was to estimate the test characteristics of the HSCL-25, in its French version (F-HSCL-25), by comparing it to the Present State Examination-9 French version (F-PSE-9) and by determining its internal validity and dimensions. Method: Outpatients from three French General Practice settings (rural, semi-rural and urban) were recruited: approximately 20,000 outpatients among 17 GPs. Two groups were formed: F-HSCL-25 ≥1.75 and F-HSCL-25 <1.75. A validated cut-off score of > 1.75 was considered to indicate a clinically relevant level of symptoms of depression and anxiety. In order to obtain two balanced groups, a different method of randomization was chosen for each group. The F-PSE-9 was randomly administered to 1 in 2 patients in the F-HSCL-25 ≥1.75 group, and to 1 in 16 in the (much larger) F-HSCL-25 <1.75 group. The diagnostic performance was assessed and the test results obtained from both groups were compared with their F-PSE-9 results. Results: Of the 1126 patients who completed the F-HCL-25, 886 joined the F-HSCL-25 <1.75 group and 240 the F-HSCL-25 ≥1.75 group. The overall prevalence of depression, using the F-HSCL-25, was 21% in these medical practices. The diagnostic performance of the F-HSCL-25 versus the F-PSE-9, the external criteria were as follows: Positive Predictive Value (PPV) 69.8%, Negative Predictive Value (NPV) 87%; Sensitivity 59.1%, and Specificity 91.4%. The Principal Component Analysis showed that F-HSCL-25 is a one-dimensional tool (anxiety and depression dimensions combined) with a Cronbach Alpha of 0.93. Conclusion: The F-HSCL-25 is an appropriate diagnostic tool for anxiety and depression in primary care in France due to its high specificity and high NPV. The HSCL-25 scale has a high eigenvalue. This pilot study will be extended throughout Europe; however, preliminary evidence suggests that the HSCL-25 is a reliable and suitable diagnostic tool for primary care. [ABSTRACT FROM AUTHOR]

Romani people are a significant minority in Europe counting about 10 million individuals scattered throughout the continent. They are a migratory group originating from Northwestern India. Their exodus from India occurred approximately 1000–1500 years ago. The migration route of the Romani people was reconstructed with the help of cultural anthropology, linguistics and historical records. Their migration made them through Central Asia, Middle East and the Caucasus region, prior to the arriving into Europe. Yet the significance of these regions, especially of the Caucasus, in Roma ancestry was a rather neglected topic. Contribution of the Caucasus and further affected regions to the ancestry of Roma was investigated based on genome-wide autosomal marker data. 158 European Roma samples and 41 populations from the Caucasus region, from Middle East, Central Asia and from South Asia were considered in our tests. Population structure and ancestry analysis algorithms were applied to investigate the relationship of Roma with these populations. Identical by descent DNA segment analyses and admixture linkage disequilibrium based tests were also applied. Our results suggest that the Caucasus region plays also a significant role in the genetic legacy of Romani people besides the main sources, Europe and South Asia, previously investigated by other population genetic studies. The Middle East and Central Asia seems slightly less important but far from negligible in connection with the sources of Roma ancestry. Our results point out that the Caucasus region and altogether the area of the Caspian and Black Seas had a significant role in the migration of Romani people towards Europe and contributed significantly to the genetic legacy of Roma rival to the European and Indian main sources. [ABSTRACT FROM AUTHOR]

Regional innovation performance is an important indicator for decision-making regarding the implementation of policies intended to support innovation. However, patterns in regional innovation structures are becoming increasingly diverse, complex and nonlinear. To address these issues, this study aims to develop a model based on a multi-output neural network. Both intra- and inter-regional determinants of innovation performance are empirically investigated using data from the 4th and 5th Community Innovation Surveys of NUTS 2 (Nomenclature of Territorial Units for Statistics) regions. The results suggest that specific innovation strategies must be developed based on the current state of input attributes in the region. Thus, it is possible to develop appropriate strategies and targeted interventions to improve regional innovation performance. We demonstrate that support of entrepreneurship is an effective instrument of innovation policy. We also provide empirical support that both business and government R&D activity have a sigmoidal effect, implying that the most effective R&D support should be directed to regions with below-average and average R&D activity. We further show that the multi-output neural network outperforms traditional statistical and machine learning regression models. In general, therefore, it seems that the proposed model can effectively reflect both the multiple-output nature of innovation performance and the interdependency of the output attributes. [ABSTRACT FROM AUTHOR]

The inferred phylogenetic relationships between organisms often depend on the molecular marker studied due to the diverse evolutionary mode and unlike evolutionary histories of different parts of the genome. Previous studies have shown conflicting patterns of differentiation of mtDNA and several nuclear markers in chamois (genus Rupicapra) that indicate a complex evolutionary picture. Chamois are mountain caprine that inhabit most of the medium to high altitude mountain ranges of southern Eurasia. The most accepted taxonomical classification considers two species, R. pyrenaica (with the subspecies parva, pyrenaica and ornata) from southwestern Europe and R. rupicapra (with the subspecies cartusiana, rupicapra, tatrica, carpatica, balcanica, asiatica and caucasica) from northeastern Europe. Phylogenies of mtDNA revealed three very old clades (from the early Pleistocene, 1.9 Mya) with a clear geographical signal. Here we analyze a set of 23 autosomal introns, comprising 15,411 nucleotides, in 14 individuals covering the 10 chamois subspecies. Introns offered an evolutionary scenario that contrasts with mtDNA. The nucleotidic diversity was 0.0013± 0.0002, at the low range of what is found in other mammals even if a single species is considered. A coalescent multilocus analysis with *BEAST indicated that introns diversified 88 Kya, in the late Pleistocene, and the effective population size at the root was lower than 10,000 individuals. The dispersal of some few migrant males should have rapidly spread trough the populations of chamois, given the homogeneity of intron sequences. The striking differences between mitochondrial and nuclear markers can be attributed to strong female philopatry and extensive male dispersal. Our results highlight the need of analyzing multiple and varied genome components to capture the complex evolutionary history of organisms. [ABSTRACT FROM AUTHOR]

The burden of disease framework facilitates the assessment of the health impact of diseases through the use of summary measures of population health such as Disability-Adjusted Life Years (DALYs). However, calculating, interpreting and communicating the results of studies using this methodology poses a challenge. The aim of the Burden of Communicable Disease in Europe (BCoDE) project is to summarize the impact of communicable disease in the European Union and European Economic Area Member States (EU/EEA MS). To meet this goal, a user-friendly software tool (BCoDE toolkit), was developed. This stand-alone application, written in C++, is open-access and freely available for download from the website of the European Centre for Disease Prevention and Control (ECDC). With the BCoDE toolkit, one can calculate DALYs by simply entering the age group- and sex-specific number of cases for one or more of selected sets of 32 communicable diseases (CDs) and 6 healthcare associated infections (HAIs). Disease progression models (i.e., outcome trees) for these communicable diseases were created following a thorough literature review of their disease progression pathway. The BCoDE toolkit runs Monte Carlo simulations of the input parameters and provides disease-specific results, including 95% uncertainty intervals, and permits comparisons between the different disease models entered. Results can be displayed as mean and median overall DALYs, DALYs per 100,000 population, and DALYs related to mortality vs. disability. Visualization options summarize complex epidemiological data, with the goal of improving communication and knowledge transfer for decision-making. [ABSTRACT FROM AUTHOR]

Medieval European urbanization presents a line of continuity between earlier cities and modern European urban systems. Yet, many of the spatial, political and economic features of medieval European cities were particular to the Middle Ages, and subsequently changed over the Early Modern Period and Industrial Revolution. There is a long tradition of demographic studies estimating the population sizes of medieval European cities, and comparative analyses of these data have shed much light on the long-term evolution of urban systems. However, the next step—to systematically relate the population size of these cities to their spatial and socioeconomic characteristics—has seldom been taken. This raises a series of interesting questions, as both modern and ancient cities have been observed to obey area-population relationships predicted by settlement scaling theory. To address these questions, we analyze a new dataset for the settled area and population of 173 European cities from the early fourteenth century to determine the relationship between population and settled area. To interpret this data, we develop two related models that lead to differing predictions regarding the quantitative form of the population-area relationship, depending on the level of social mixing present in these cities. Our empirical estimates of model parameters show a strong densification of cities with city population size, consistent with patterns in contemporary cities. Although social life in medieval Europe was orchestrated by hierarchical institutions (e.g., guilds, church, municipal organizations), our results show no statistically significant influence of these institutions on agglomeration effects. The similarities between the empirical patterns of settlement relating area to population observed here support the hypothesis that cities throughout history share common principles of organization that self-consistently relate their socioeconomic networks to structured urban spaces. [ABSTRACT FROM AUTHOR]

Analysis of the proportion of immature skeletons recovered from European prehistoric cemeteries has shown that the transition to agriculture after 9000 BP triggered a long-term increase in human fertility. Here we compare the largest analysis of European cemeteries to date with an independent line of evidence, the summed calibrated date probability distribution of radiocarbon dates (SCDPD) from archaeological sites. Our cemetery reanalysis confirms increased growth rates after the introduction of agriculture; the radiocarbon analysis also shows this pattern, and a significant correlation between both lines of evidence confirms the demographic validity of SCDPDs. We analyze the areal extent of Neolithic enclosures and demographic data from ethnographically known farming and foraging societies and we estimate differences in population levels at individual sites. We find little effect on the overall shape and precision of the SCDPD and we observe a small increase in the correlation with the cemetery trends. The SCDPD analysis supports the hypothesis that the transition to agriculture dramatically increased demographic growth, but it was followed within centuries by a general pattern of collapse even after accounting for higher settlement densities during the Neolithic. The study supports the unique contribution of SCDPDs as a valid demographic proxy for the demographic patterns associated with early agriculture. [ABSTRACT FROM AUTHOR]

Providing an underutilized source of information for paleoenvironmental reconstructions, birds are rarely used to infer paleoenvironments despite their well-known ecology and extensive Quaternary fossil record. Here, we use the avian fossil record to investigate how Western Palearctic bird assemblages and species ranges have changed across the latter part of the Pleistocene, with focus on the links to climate and the implications for vegetation structure. As a key issue we address the full-glacial presence of trees in Europe north of the Mediterranean region, a widely debated issue with evidence for and against emerging from several research fields and data sources. We compiled and analyzed a database of bird fossil occurrences from archaeological sites throughout the Western Palearctic and spanning the Saalian-Eemian-Weichselian stages, i.e. 190,000–10,000 years BP. In general, cold and dry-adapted species dominated these late Middle Pleistocene and Late Pleistocene fossil assemblages, with clear shifts of northern species southwards during glacials, as well as northwards and westwards shifts of open-vegetation species from the south and east, respectively and downwards shifts of alpine species. A direct link to climate was clear in Northwestern Europe. However, in general, bird assemblages more strongly reflected vegetation changes, underscoring their usefulness for inferring the vegetation structure of past landscapes. Forest-adapted birds were found in continuous high proportions throughout the study period, providing support for the presence of trees north of the Alps, even during full-glacial stages. Furthermore, the results suggest forest-dominated but partially open Eemian landscapes in the Western Palearctic, including the Northwestern European subregion. [ABSTRACT FROM AUTHOR]