Abstract
OBJECTIVE: To evaluate endothelial dysfunction and microvascular damage in secondary Raynaud Phenomenon (SRP) and Systemic sclerosis (SSc)-associated patients as possible predictors of ischemic fingertip digital ulcers (DU) in a 3-year clinical follow-up. METHODS: Flow-mediated dilatation (FMD), nailfold videocapillaroscopy (NVC), endothelin-1 (ET-1) and asymmetric dimethylarginine (ADMA) were analysed in a 3-year observational cohort study of 77 SRP patients with systemic sclerosis. The primary outcome was the occurrence of a new DU. RESULTS: Risk factors for DU at baseline were low FMD% (p < 0.001), NVC pattern (p < 0.001), high microangiopathy evolution score (MES) (p < 0.001), increased ET-1 (p < 0.001) and increased ADMA serum levels (p = 0.001). Median time to the occurrence of a new DU was 4.50 (1.25-16.25) months. The risk factors for the occurrence of at least one new DU episode in follow-up included a history of at least one DU before enrolment (p < 0.001), autoantibody anti-scleroderma-70 (p = 0.012), NVC late pattern (p < 0.001), high MES score (p < 0.001), low FMD% (p < 0.001) and increased ET-1 serum levels (p < 0.001). We used univariate Cox regression analysis to show that FMD > 9.41% (HR: 0.37 95% CI: 0.14-0.99) and ET-1 > 11.85 pmol/L (HR: 3.81 95% CI: 1.41-10.26) and NVC (HR: 2.29 95% CI: 0.97-5.38) were predictors of DU recurrence. In terms of first DU event in naive DU patients at baseline, late NVC pattern (HR: 12.66 95% CI: 2.06-77.89) and MES score (HR: 1.693 95% CI: 1.257-2.279) were independent predictors. CONCLUSION: This study identified endothelium dysfunction biomarkers (FMD and ET-1) and severe microvascular damage in NVC as strong predictors of new DU in SSc patients.

Abstract

Abstract

Abstract
In recent years we witnessed the arrival of new trends, such as server virtualization and cloud services, an increasing number of mobile devices and online contents, leading the networking industry to deliberate about how traditional network architectures can be adapted or even deciding if a new perspective for them should be taken. SDN (Software-Defined Networking) emerged under this framing, opening a road for new developments due to the centralized logic control and view of the network, the decoupling of data and control planes, and the abstraction of the underlying network infrastructure from the applications. Although firstly oriented to packet switching, network measurements have also emerged as one promising field for SDN, as its flexibility enables programmable measurements, allowing a SDN controller to manage measurement tasks concurrently at multiple spatial and temporal scales. In this context, this paper is focused on exploring the SDN architecture and components for supporting the flexible selection and configuration of network monitoring tasks that rely on the use of traffic sampling. The aim is to take advantage of the integrated view of SDN controllers to apply and configure appropriate sampling techniques in network measurement points according to the requirements of specific measurement tasks. Through SDN, flexible and service-oriented configuration of network monitoring can be achieved, allowing also to improve the trade-off between accuracy and overhead of the monitoring process. In this way, this study, examining relevant SDN elements and solutions for deploying this monitoring paradigm, provides useful insights to enhance the programmability and efficiency of sampling-based network monitoring. © 2017, Springer International Publishing AG.

Abstract
The problem of Label Ranking is receiving increasing attention from several research communities. The algorithms that have been developed/adapted to treat rankings of a fixed set of labels as the target object, including several different types of decision trees (DT). One DT-based algorithm, which has been very successful in other tasks but which has not been adapted for label ranking is the Random Forests (RF) algorithm. RFs are an ensemble learning method that combines different trees obtained using different randomization techniques. In this work, we propose an ensemble of decision trees for Label Ranking, based on Random Forests, which we refer to as Label Ranking Forests (LRF). Two different algorithms that learn DT for label ranking are used to obtain the trees. We then compare and discuss the results of LRF with standalone decision tree approaches. The results indicate that the method is highly competitive.

Abstract
The construction of earthen embankment terraces in the Douro Region raises a set of problems related to hydrological processes. The main objective of this study is the evaluation of the spatial variation of electrical resistivity in agriculture terraces at Douro valley (Portugal). To achieve this objective, two variables are analysed, the soil electrical resistivity and the flow direction algorithm. In a field survey we recorded 13 electrical resistivity profiles. The contributing area was calculated with the algorithms D infinity (Deterministic Infinity Flow) and MFD (Multiple Flow Direction) and the results are the base of the internal runoff modelling, both supported by the digital elevation model with a spatial resolution of 1m2. A correlation between the spatial variation of the soil electrical resistivity represented by the standard deviation of the electrical resistivity for each profile and the average value of the contributing area coincident with each profile was established. The electrical resistivity standard deviation seems to be moderately well correlated according to the D infinity algorithm at about 1m of depth, and it has a good correlation at 1,5m to 2m of depth with the MFD algorithm. Taken together, the results show a significant positive statistical correlation between the electrical resistivity standard deviation and the contributing areas (MFD and D infinity) depending on the soil depth.