Abstract
Machine Learning and Data Mining research strongly depend on the quality and quantity of the real world datasets for the evaluation stages of the developing methods. In the context of the emerging Online Multi-Target Regression and Multi-Label Classification methodologies, datasets present new characteristics that require specific testing and represent new challenges. The first difficulty found in evaluation is the reduced amount of examples caused by data damage, privacy preservation or high cost of acquirement. Secondly, few data events of interest such as data changes are difficult to find in the datasets of specific domains, since these events naturally scarce. For those reasons, this work suggests a method of producing synthetic datasets with desired properties(number of examples, data changes events, ... ) for the evaluation of Multi-Target Regression and Multi-Label Classification methods. These datasets are produced using First Principle Models which give more realistic and representative properties such as real world meaning ( physical, financial, ... ) for the outputs and inputs variables. This type of dataset generation can be used to produce infinite streams and to evaluate incremental methods such as online anomaly and change detection. This paper illustrates the use of synthetic data generation through two showcases of data changes evaluation.

Abstract

Abstract
This paper describes the application of the Opt-aiNet algorithm to the reconfiguration problem of distribution systems considering variable demand levels. The Opt-aiNet algorithm is an optimization technique inspired in the immunologic bio system and it aims at reproducing the main properties and functions of this system. The reconfiguration problem of distribution networks with variable demands is a complex problem that aims at identifying the most adequate radial topology of the network that complies with all technical constraints in every demand level while minimizing the cost of power losses along an extended operation period. This work includes results of the application of the Opt-aiNet algorithm to distribution systems with 33, 84, 136 and 417 buses. These results demonstrate the robustness and efficiency of the proposed approach.

Abstract
This paper presents Enki, a web-based IDE that integrates several pedagogical tools designed to engage students in learning programming languages. Enki achieves this goal (1) by sequencing educational resources, either expository or evaluative, (2) by using gamification services to entice students to solve activities, (3) by promoting social interaction and (4) by helping students with activities, providing feedback on submitted solutions. The paper describes Enki, its concept and architecture, details its design and implementation, and covers also its validation.

Abstract
Ongoing declines in biodiversity caused by global environmental changes call for adaptive conservation management, including the assessment of habitat suitability spatiotemporal dynamics potentially affecting species persistence. Remote sensing (RS) provides a wide-range of satellite-based environmental variables that can be fed into species distribution models (SDMs) to investigate species-environment relations and forecast responses to change. We address the spatiotemporal dynamics of species' habitat suitability at the landscape level by combining multi-temporal RS data with SDMs for analysing inter-annual habitat suitability dynamics. We implemented this framework with a vulnerable plant species (Veronica micrantha), by combining SDMs with a time-series of RS-based metrics of vegetation functioning related to primary productivity, seasonality, phenology and actual evapotranspiration. Besides RS variables, predictors related to landscape structure, soils and wildfires were ranked and combined through multi-model inference (MMI). To assess recent dynamics, a habitat suitability time-series was generated through model hindcasting. MMI highlighted the strong predictive ability of RS variables related to primary productivity and water availability for explaining the test-species distribution, along with soil, wildfire regime and landscape composition. The habitat suitability time-series revealed the effects of short-term land cover changes and inter-annual variability in climatic conditions. Multi-temporal SDMs further improved predictions, benefiting from RS time-series. Overall, results emphasize the integration of landscape attributes related to function, composition and spatial configuration for improving the explanation of ecological patterns. Moreover, coupling SDMs with RS functional metrics may provide early-warnings of future environmental changes potentially impacting habitat suitability. Applications discussed include the improvement of biodiversity monitoring and conservation strategies.

Abstract
In this work we identify a seminal design guideline that prevents current Full-Duplex (FD) MAC protocols to scale the FD capacity gain (i.e. 2× the half-duplex throughput) in single-cell Wi-Fi networks. Under such guideline (referred to as 1-1), a MAC protocol attempts to initiate up to two simultaneous transmissions in the FD bandwidth. Since in single-cell Wi-Fi networks MAC performance is bounded by the PHY layer capacity, this implies gains strictly less than 2× over half-duplex at the MAC layer. To face this limitation, we argue for the 1:N design guideline. Under 1:N, FD MAC protocols 'see' the FD bandwidth through N>1 orthogonal narrow-channel PHY layers. Based on theoretical results and software defined radio experiments, we show the 1:N design can leverage the Wi-Fi capacity gain more than 2× at and below the MAC layer. This translates the denser modulation scheme incurred by channel narrowing and the increase in the spatial reuse factor enabled by channel orthogonality. With these results, we believe our design guideline can inspire a new generation of Wi-Fi MAC protocols that fully embody and scale the FD capacity gain. © 2016 IEEE.