Abstract
Object oriented decomposition is the most successful decomposition strategy used nowadays. But a single decomposition strategy cannot capture all aspects of a concept. Roles have been successfully used to model the different views a concept may provide but, despite this, roles have not been used as building blocks. Roles are mostly used to extend objects at runtime. In this paper we propose roles as a way to compose classes that provides a modular way of capturing and reusing those aspects that fall outside a concept's main purpose, while being close to the OO approach. We present how roles can be made modular and reusable. We also show how we can use roles to compose classes using JavaStage, a java extension that support roles To validate our approach we developed generic and reusable roles for the Gang of Four patterns. We were able to develop reusable roles for 10 out of 23 patterns, which is a good outcome.

Abstract
Roles are not a new concept, but they have been used in two different ways: as modeling concepts in a static view and as instance extensions in a dynamic view. For these views only the dynamic offers supporting languages. The static view, although proving the utility of roles in modeling, does not offer a programming language that allows developers to use roles all the way from modeling to programming. We try to overcome this by presenting our role language JavaStage, based on the Java language. We do this by designing and implementing a simple framework and then compare the results with its OO equivalent. Our results show that static roles are in fact useful when used in code and that JavaStage features expand role reuse.

Abstract

Abstract
Depression is expected to be the disorder with the highest disease burden in high-income countries by the year 2030. ICT4Depression (ICT4D) is a European FP7 project, which aims to contribute to the alleviation of this burden by making use of depression treatment and ICT innovations. In this project we developed an ICT-based system for use in primary care that aims to improve access as well as actual care delivery for depressed adults. Innovative technologies within the ICT4D system include 1) flexible self-help treatments for depression, 2) automatic assessment of the patient using mobile phone and web-based communication 3) wearable biomedical sensor devices for monitoring activities and electrophysiological indicators, 4) computational methods for reasoning about the state of a patient and the risk of relapse (reasoning engine) and 5) a flexible system architecture for monitoring and supporting people using continuous observations and feedback via mobile phone and the web. The general objective of the ICT4D project is to test the feasibility and acceptability of the ICT4D system within a pilot study in the Netherlands and in Sweden during 2012 and 2013. © 2012 Interactive Media Institute and IOS Press.

Abstract
There has been increasing interest in pattern classification methods and neuroimaging studies using permutation tests to estimate the statistical significance of a classifier (p-value). Permutation tests usually use the test error as a dataset statistic to estimate the p-value(s) by measuring the dissimilarity between two or more populations. Using the test error as a dataset statistic; however, may camouflage the lowest recognizable classes, and the resulting p-value will be biased toward better values (usually lower values) because of the highly recognizable classes; thus, lower p-values could sometimes be the result of undercoverage. In this study, we investigate this problem and propose the implementation of permutation tests based on a per-class test error as a dataset statistic. We also propose a model that is based on partially scrambling the testing samples (in this model, the training samples are not scrambled) when computing the non-permuted statistic in order to judge the p-value's tolerance and to draw conclusions regarding, which permutation test procedures are more reliable. For the same purpose, we propose another model that is based on chance-level shifting of the permuted statistic. We tested these two proposed models on functional magnetic resonance imaging data that were collected while human subjects responded to visual stimulation paradigms, and our results showed that these models can aid in determining, which permutation test procedure is superior. We also found that permutation tests that use a per-class test error as a dataset statistic are more reliable in addressing the null hypothesis that all classes in the problem domain are drawn from the same distribution.

Abstract
The large amount of pages in Websites is a problem for users who waste time looking for the information they really want. Knowledge about users' previous visits may provide patterns that allow the customization of the Website. This concept is known as Adaptive Website: a Website that adapts itself for the purpose of improving the user's experience. Some Web Mining algorithms have been proposed for adapting a Website. In this paper, a recommender system using agents with two different algorithms (associative rules and collaborative filtering) is described. Both algorithms are incremental and work with binary data. Results show that this multi-agent approach combining different algorithms is capable of improving user's satisfaction.