Abstract
The goal of ensemble regression is to combine several models in order to improve the prediction accuracy in learning problems with a numerical target variable. The process of ensemble learning can be divided into three phases: the generation phase, the pruning phase, and the integration phase. We discuss different approaches to each of these phases that are able to deal with the regression problem, categorizing them in terms of their relevant characteristics and linking them to contributions from different fields. Furthermore, this work makes it possible to identify interesting areas for future research.

Abstract
Being political entrepreneurship an issue that has received a high distinction in literature, its empirical evidence is scarce largely derived from limitations of operationalizing the concept. In this article we operationalize local political entrepreneurship considering three main dimensions: 1) fundraising, 2) the possession and/or construction of infrastructure to support business and social activities, and 3) the supply of support activities and services required for activities business and social. Based on information gathered from 108 Portuguese municipalities, we quantified the overall political entrepreneurship and in the three dimensions mentioned presenting a ranking of municipalities and the corresponding NUTS III regions. Additionally, through the estimation of econometric specifications, we conclude that the local political entrepreneurship has a positive and significant but indirect impact, through the human capital of employed population, on the creation of new businesses, particularly the knowledge-intensive ones. In other words, the impact of entrepreneurial actions by the local political authorities is much higher the greater the proportion of the employed population with higher education in the region. On this basis, it is not enough that local authorities are 'entrepreneurial'; in addition to the public actions of fundraising, construction of infrastructures and promotion of activities for business promotion, one needs to ensure that the region is endowed with an adequate level of absorptive capacity, i.e., investment in human capital and R&D activities.

Abstract
This paper describes a Dynamic Model of the electricity sector that can be used to simulate the evolution of some key variables on the long term, namely the evolution of the electricity price, of the demand and of the capacity factors of the technologies in the generation mix. This model can be used in different ways and by several agents, for instance to estimate the impact on the electricity price of the increasing presence of renewable power stations, namely using wind power and PV systems. In several countries these stations are paid feed-in tariffs with a fixed price but in some cases this scheme is under discussion and there are opinions that payments determined by the market price are more adequate and would bring fewer costs to final consumers. Such a change has to be carefully evaluated given that the presence of renewable stations bidding at an infra marginal price will affect the price itself. The model described in this paper can be used in a profitable way both by governmental agencies when preparing or studying alternative remuneration schemes to renewable stations or by promoters themselves to get more insight to the profitability of their investments, namely if the fixed feed-in tariffs in force in several countries are changed.

Abstract
The economic sector of contact centers is growing by more than 8% a year. It is a multidisciplinary area in which information systems are decisive to organizations' success. Contact Centers' Information Systems deal with real time requisites and critical business information. A theory-building research shows a framework with 12 key design factors to consider, which managers might use to develop projects and researchers may adopt for further investigation in the area of Contact Center design. This work intends to provide a valuable link between the research community and practitioners in industry.

Abstract
The energy storage system (ESS), as well as its efficient management, represents a key factor for the success of electric vehicles. Due to well-known technological constraints in ESSs, there has been a growing interest in combining various types of energy sources with complementary features. Among the many possible combinations, our interest here lies in the hybridization of batteries and supercapacitors (SCs) with an active parallel arrangement, i.e., the sources are connected to the direct current (dc) bus through two bidirectional dc-dc converters (step-up). Based on this ESS topology, a robust dc-link controller is employed to regulate the dc-bus voltage and track the SCs current in spite of uncertainties in the system. For this purpose, we start by showing that the converters' uncertainty, e. g., the powertrain load, can be modeled as a convex polytope. The dc-link controller is then posed as a robust linear-quadratic regulator problem and, by exploring the convex polytope, converted in a linear matrix inequalities framework, which can efficiently be solved by numerical means. Finally, the operation envelope of the controller is extended by scheduling the gains according to the energy sources' voltages, which is an important feature to cope with the voltage variations in the SCs. To analyze the performance of the control architecture, a reduced-scale prototype was built. The experimental results show that, compared with the nonrobust and non-gain-scheduled controllers, the proposed dc-link controller offers better transient response and robustness to disturbances. Furthermore, the global performance of the controller is evaluated during certain driving cycles.

Abstract
Data streams are usually characterized by changes in the underlying distribution generating data. Therefore algorithms designed to work with data streams should be able to detect changes and quickly adapt the decision model. Rules are one of the most interpretable and flexible models for data mining prediction tasks. In this paper we present the Adaptive Very Fast Decision Rules (AVFDR), an on-line, any-time and one-pass algorithm for learning decision rules in the context of time changing data. AVFDR can learn ordered and unordered rule sets. It is able to adapt the decision model via incremental induction and specialization of rules. Detecting local drifts takes advantage of the modularity of rule sets. In AVFDR, each individual rule monitors the evolution of performance metrics to detect concept drift. AVFDR prunes rules that detect drift. This explicit change detection mechanism provides useful information about the dynamics of the process generating data, faster adaption to changes and generates compact rule sets. The experimental evaluation shows this method is able to learn fast and compact rule sets from evolving streams in comparison to alternative methods. © 2012 Springer-Verlag.