Abstract
A cancer registry is a standardized tool to produce population-based data on cancer incidence and survival. Cancer registries can retrieve and store information on all cancer cases occurring in a defined population. The main sources of data on cancer cases usually include: treatment and diagnostic facilities (oncology centres or hospital departments, pathology laboratories, or imaging facilities etc.) and the official territorial death registry. The aim of this paper is to evaluate the north regional cancer registry (RORENO) of Portugal using a qualitative research. We want to characterize: the main functionalities and core processes, team involved, different healthcare institutions in the regional network and an identification of issues and potential improvements. RORENO links data of thirteen-two healthcare institutions and is responsible for the production of cancer incidence and survival report for this region. In our semi-structure interviews and observation of RORENO we identified a serious problem due to a lack of an automatic integration of data from the different sources. Most of the data are inserted manually in the system and this implies an extra effort from the RORENO team. At this moment RORENO team are still collecting data from 2011. In a near future it is crucial to automatize the integration of data linking the different healthcare institutions in the region. However, it is important to think which functionalities this system should give to the institutions in the network to maximize the engagement with the project. More than a database this should be a source of knowledge available to all the collaborative oncologic network.

Abstract
Credit-card fraud leads to billions of dollars in losses for online merchants. With the development of machine learning algorithms, researchers have been finding increasingly sophisticated ways to detect fraud, but practical implementations are rarely reported. We describe the development and deployment of a fraud detection system in a large e-tail merchant. The paper explores the combination of manual and automatic classification, gives insights into the complete development process and compares different machine learning methods. The paper can thus help researchers and practitioners to design and implement data mining based systems for fraud detection or similar problems. This project has contributed not only with an automatic system, but also with insights to the fraud analysts for improving their manual revision process, which resulted in an overall superior performance.

Abstract
The fast economic growth of China along the last two decades has created a strong impact on the environment. The occurrence of heavy haze pollution days is the most visible effect. Although many researchers have studied such problem, a high number of spatio-temporal limitations in the recent studies were identified. From our best knowledge the long trends of PM2.5 concentrations were not fully investigated in China, in particular the year-to-year trends and the seasonal and daily cycles. Therefore, in this work the PM2.5 concentrations collected from automatic monitors from five urban sites located in megacities with different climatic zones in China were analysed: Beijing (40 degrees N), Chengdu (31 degrees N), Guangzhou (23 degrees N), Shanghai (31 degrees N) and Shenyang (43 degrees N). For an inter-comparison a meta-analysis was carried out. An evaluation conducted since 1999 demonstrates that PM2.5 concentrations have been reduced until 2008, period which match with the occurrence of the Olympic Games. However, a seasonal analysis highlight that such decrease occurs mostly during warmer seasons than cold seasons. During winter PM2.5 concentrations are typically 1.3 to 2.7 higher than in summer. The average daily cycle shows that the lowest and highest PM2.5 concentrations often occurs in the afternoon and evening hours respectively. Such daily variations are mostly driven by the daily variation of the boundary layer depth and emissions. Although the PM2.5 levels have showing signs of improvement, even during the warming season the values are still too high in comparison with the annual environmental standards of China (35 mu g m(-3)). Moreover, during cold seasons the north regions have values twice higher than this limit. Thus, to fulfil these standards the governmental mitigation measures need to be strongly reinforced in order to optimize the daily living energy consumption, primarily in the north regions of China and during the winter periods.

Abstract
The Brazilian power generation sector faces a paradigm change driven by, on one hand, a shift from a hydropower dominated mix and, on the other hand, international goals for reducing greenhouse gas emissions. The objective of this work is to evaluate five scenarios for the Brazilian power sector until 2050 using a multi-criteria decision analysis tool. These scenarios include a baseline trend and low carbon policy scenarios based on carbon taxes and carbon emission limits. To support the applied methodology, a questionnaire was elaborated to integrate the perceptions of experts on the scenario evaluation process. Considering the results from multi-criteria analysis, scenario preference followed the order of increasing share of renewables in the power sector. The preferable option for the future Brazilian power sector is a scenario where wind and biomass have a major contribution. The robustness of the multi-criteria tool applied in this study was tested by a sensitivity analysis. This analysis demonstrated that, regardless of the respondents' preferences and backgrounds, scenarios with higher shares of fossil fuel sources are the least preferable option, while scenarios with major contributions from wind and biomass are the preferable option to supply electricity in Brazil through 2050.

Abstract
---

Abstract