Abstract
In some classification tasks, such as those related to the automatic building and maintenance of text corpora, it is expensive to obtain labeled instances to train a classifier. In such circumstances it is common to have massive corpora where a few instances are labeled (typically a minority) while others are not. Semi-supervised learning techniques try to leverage the intrinsic information in unlabeled instances to improve classification models. However, these techniques assume that the labeled instances cover all the classes to learn which might not be the case. Moreover, when in the presence of an imbalanced class distribution, getting labeled instances from minority classes might be very costly, requiring extensive labeling, if queries are randomly selected. Active learning allows asking an oracle to label new instances, which are selected by criteria, aiming to reduce the labeling effort. D-Confidence is an active learning approach that is effective when in presence of imbalanced training sets. In this paper we evaluate the performance of d-Confidence in comparison to its baseline criteria over tabular and text datasets. We provide empirical evidence that d-Confidence reduces label disclosure complexity-which we have defined as the number of queries required to identify instances from all classes to learn-when in the presence of imbalanced data. © 2012 The Brazilian Computer Society.

Abstract
This paper performs a survey on the state of the art in the security and access controls mechanisms employed by multi agent systems as a form of data transaction, with special emphasis on their security threats and how these issues are currently being tackled for multi-agent systems. In this paper we provide a brief summary of current approaches and provide comments on their relevance, implementation and safety assumptions behind their safe operation on real systems with sensitive data. We also take special notice on the security issues approached by each one of the analysed proposals and the sometimes ingenious ways proposed by different authors to solve them. We also pay special attention to issues related with the authentication, confidentiality, integrity and access control mechanisms employed to minimise security incidents. Nowadays multi-agent systems are increasingly being researched and deployed to solve complex interoperability issues that normally occur between different institutions. The final result of this survey is a summary aggregation of knowledge about security and access control practices for multi agent models and implementations.

Abstract
Exploiting the benefit that one can get from large scale integration of microgeneration (mu G) requires the adoption of different control strategies at different distribution network levels. In that sense, MicroGrid (MG) and Multi MicroGrids (MMG) concepts may be seen as a way to facilitate integration of higher levels of mu G into the system. This paper describes new control and management functionalities used to take profit from the presence of microgeneration and active loads under MG and MMG concepts. MultiCriteria Decision Aid (MCDA) techniques are used to capture decision maker's preferences in evaluating the impact of MG and MMG concepts deployment and help selecting the best development strategy.

Abstract
This paper is concerned with the protection of wind energy systems against the direct effects of lightning. As wind power generation undergoes rapid growth, lightning damages involving wind turbines have come to be regarded as a serious problem. Nevertheless, very few studies exist yet in Portugal regarding lightning protection of wind energy systems using numerical codes. A new case study is presented in this paper, based on a wind turbine with an interconnecting transformer, for the analysis of transient phenomena due to a direct lightning strike to the blade. Comprehensive simulation results are provided by using models of the Restructured Version of the Electro-Magnetic Transients Program (EMTP), and conclusions are duly drawn.

Abstract
The purpose of this study was to assess the effects of a six-months' nutrition program, delivered and taught by classroom teachers with in-service nutrition training, on the prevention of overweight and obesity among children in grades 1 to 4. In this randomized trial, four hundred and sixty four children from seven elementary schools were allocated to a nutrition educational program delivered by their own teachers. Intervened teachers had 12 sessions of three hours each with the researchers throughout six months, according to the topics nutrition and healthy eating, the importance of drinking water and healthy cooking activities. After each session, teachers were encouraged to develop activities in class focused on the learned topics. Sociodemographic, anthropometric, dietary, and physical activity assessments were performed at baseline and at the end of the intervention. In the intervention group the increase in Body Mass Index (BMI) z-score was significantly lower than in the control group (p = 0.009); fewer proportion of children became overweight in the intervened group compared with the control (5.6% vs. 18.4%; p = 0.037). Our study provides further support to decrease the overweight epidemic, involving classroom teachers in a training program and making them dedicated interventionists.

Abstract
Universities are continually pursuing ways to adapt their educational practices, looking to develop new learning culture that encourages creativity and active engagement. The adoption of augmented lecture podcast scenarios, based on a content-centric paradigm, looks a promising way to reach a new level of interactivity. This paper analyzes the main advantages provided by this innovative approach comparing it with the traditional podcasting developing model. Additionally, it proposes possible business models based on content distribution and augmentation, to enable wider exploitation of the approach within university environments. © 2012 Infonomics Society.