Abstract
Metalearning is a subfield of machine learning with special propensity for dynamic and complex environments, from which it is difficult to extract predictable knowledge. The field of study of this work is the electricity market, which due to the restructuring that recently took place, became an especially complex and unpredictable environment, involving a large number of different entities, playing in a dynamic scene to obtain the best advantages and profits. This paper presents the development of a metalearner, applied to the decision support of electricity markets' negotiation entities. The proposed metalearner takes advantage on several learning algorithms implemented in ALBidS, an adaptive learning system that provides decision support to electricity markets' participating players. Using the outputs of each different strategy as inputs, the metalearner creates its own output, considering each strategy with a different weight, depending on its individual quality of performance. The results of the proposed method are studied and analyzed using MASCEM - a multi-agent electricity market simulator that models market players and simulates their operation in the market. This simulator provides the chance to test the metalearner in scenarios based on real electricity markets' data.

Abstract
Traffic sampling techniques are crucial and extensively used to assist network management tasks. Nevertheless, combining accurate network parameters' estimation and flexible lightweight measurements is an open challenge. In this context, this paper proposes a self-adaptive sampling technique, based on linear prediction, which allows to reduce significantly the measurement overhead, while assuring that sampled traffic reflects the statistical characteristics of the global traffic under analysis. The technique is multiadaptive as several parameters are considered in the dynamic configuration of the traffic selection process. The devised test scenarios aim at exploring the proposed sampling technique ability to join accurate network estimates to reduced overhead, using throughput as reference parameter. The evaluation results, obtained resorting to real traffic traces representing wired and wireless aggregated traffic scenarios and actual network services, prove that the simplicity, flexibility and self-adaptability of this technique can be successfully explored to improve network measurements efficiency over distinct traffic conditions. For optimization purposes, this paper also includes a study of the impact of varying the order of prediction, i.e., of considering different degrees of past memory in the self-adaptive estimation mechanism. The significance of the obtained results is demonstrated through statistical benchmarking.

Abstract

Abstract
In Computed Tomography (CT), bone segmentation is considered an important step to extract bone parameters, which are frequently useful for computer-aided diagnosis, surgery and treatment of many diseases such as osteoporosis. Consequently, the development of accurate and reliable segmentation techniques is essential, since it often provides a great impact on quantitative image analysis and diagnosis outcome. This chapter presents an automated multistep approach for bone segmentation in volumetric CT datasets. It starts with a three-dimensional (3D) watershed operation on an image gradient magnitude. The outcome of the watershed algorithm is an over-partioning image of many 3D regions that can be merged, yielding a meaningful image partitioning. In order to reduce the number of regions, a merging procedure was performed that merges neighbouring regions presenting a mean intensity distribution difference of ±15%. Finally, once all bones have been distinguished in high contrast, the final 3D bone segmentation was achieved by selecting all regions with bone fragments, using the information retrieved by a threshold mask. The bones contours were accurately defined according to the watershed regions outlines instead of considering the thresholding segmentation result. This new method was tested to segment the rib cage on 185 CT images, acquired at the São João Hospital of Porto (Portugal) and evaluated using the dice similarity coefficient as a statistical validation metric, leading to a coefficient mean score of 0.89. This could represent a step forward towards accurate and automatic quantitative analysis in clinical environments and decreasing time-consumption, user dependence and subjectivity.

Abstract
Emergency evacuation plans and evacuation drills are mandatory in public buildings in many countries. Their importance is considerable when it comes to guarantee safety and protection during a crisis. However, sometimes discrepancies arise between the goals of the plan and its outcomes, because people find it hard to take them very seriously, or due to the financial and time resources required. Serious games are a possible solution to tackle this problem. They have been successfully applied in different areas such as health care and education, since they can simulate an environment/task quite accurately, making them a practical alternative to real-life simulations. This paper presents a serious game developed using Unity3D to recreate a virtual fire evacuation training tool. The prototype application was deployed which allowed the validation by user testing. A sample of 30 individuals tested the evacuating scenario, having to leave the building during a fire in the shortest time possible. Results have shown that users effectively end up learning some evacuation procedures from the activity, even if only to look for emergency signs indicating the best evacuation paths. It was also evidenced that users with higher video game experience had a significantly better performance.

Abstract
High-maturity software development processes, making intensive use of metrics and quantitative methods, such as the Personal Software Process (PSP) and the Team Software Process (TSP), can generate a significant amount of data that can be periodically analyzed to identify performance problems, determine their root causes and devise improvement actions. Currently, there are several tools that automate data collection and produce performance charts for manual analysis in the context of the PSP/TSP, but practically no tool support exists for automating the data analysis and the recommendation of improvement actions. Manual analysis of this performance data is problematic because of the large amount of data to analyze and the time and expertise required. Hence, we propose in this paper a performance model and a tool (named PSP PAIR) to automate the analysis of performance data produced in the context of the PSP, namely, identify performance problems and their root causes, and recommend improvement actions. The work presented is limited to the analysis of the time estimation performance of PSP developers, but is extensible to other performance indicators and development processes.