Abstract
Most current customer baseline load (CBL) estimation methods for incentive-based demand response (DR) rely heavily on historical data and are unable to adapt to the cases when the load patterns (LPs) in the DR event day are not similar enough to those in non-DR days. After the error generation mechanism of current methods is revealed, a synchronous pattern matching principle-based residential CBL estimation approach without historical data requirement is proposed. All customers are split into DR and CONTROL group, including DR participants and non-DR customers, respectively. First, all CONTROL group customers are clustered into several non-overlapping clusters according to LPs similarity in the DR event day. Second, each DR participant is matched to the most similar cluster in the CONTROL group according to the similarity between its load curve segments in DR event day, excluding DR part and cluster centroids. Third, the CBL of each DR participant is estimated with an optimized weight combination method using the load data within the DR event period of all the customers in the very matching cluster in the CONTROL group. A comparison with five well-known CBL estimation methods using a dataset of 736 residential customers indicates that the proposed approach has better overall performance than other current CBL estimation methods.

Abstract
Although cross-validation is a standard procedure for performance evaluation, its joint application with oversampling remains an open question for researchers farther from the imbalanced data topic. A frequent experimental flaw is the application of oversampling algorithms to the entire dataset, resulting in biased models and overly-optimistic estimates. We emphasize and distinguish overoptimism from overfitting, showing that the former is associated with the cross-validation procedure, while the latter is influenced by the chosen oversampling algorithm. Furthermore, we perform a thorough empirical comparison of well-established oversampling algorithms, supported by a data complexity analysis. The best oversampling techniques seem to possess three key characteristics: use of cleaning procedures, cluster-based example synthetization and adaptive weighting of minority examples, where Synthetic Minority Oversampling Technique coupled with Tomek Links and Majority Weighted Minority Oversampling Technique stand out, being capable of increasing the discriminative power of data.

Abstract
Scheduling jobs on a set of identical parallel machines using efficient heuristics when the objective is to minimize total weighted squared tardiness is considered. Two efficient heuristics and an improvement procedure are presented for the problem. These heuristics and other heuristics are tested using problem sets that represent a variety of conditions. The results show that one of the heuristics consistently performs better than the other heuristics tested. It is also shown how these heuristics can be incorporated into other procedures such as the existing Lagrangian relaxation procedure or meta-heuristics to obtain improved solutions for medium sized problems.

Abstract
As organizations are entering social media, determining their current strategy will allow to combine monitoring and benchmarking methods to foster the identification of opportunities and threats, which can serve as inputs for the evaluation of social media strategies' and eventual readjustments, and a subsequent efficiency measurement. Higher Educational Institutions (HEI) are not different from other organizations in which concerns these problems. To address these challenges, we propose an automatic procedure to assess the posting behavior and strategy identification for each higher educational institution. We used a sample of the 10-top worldwide ranked educational institutions in this study and collected the posts from their official Facebook pages during an entire school year. Our study was conducted on the frequency and intensity of publications by universities, which included an analysis of the number of responses to 'posts' over time in the form of 'shares'. Finally, the content of the posts was analyzed according to the topics covered in the messages. This process allowed us to identify the editorial areas that each university uses the most and in which are more active. © 2018 The Authors. Published by Elsevier Ltd..

Abstract
Oil spill incidents in the sea or harbors occur with some regularity during exploration, production, and transport of petroleum products. In order to mitigate the impact of the oil spill in the marine life, immediate, safety, effective and ecofriendly actions must be taken. Autonomous vehicles can assume an important contribution by establishing a cooperative and coordinated intervention. This paper presents the development of a path planning control-law methods for an autonomous surface vehicle (ASV) being able to contour the oil spill while is deploying microorganisms and nutrients (bioremediation) capable of mitigating and contain the oil spill spread with the collaboration of a UAV vehicle. An oil spill simulation scenario was developed in Gazebo to support the evaluation of the cooperative actions between the ASV and UAV and to infer the ASV path planning for each one of the proposed control-law methods.

Abstract
One of the main difficulties in the use of deep learning strategies in medical contexts is the training set size. While these methods need large annotated training sets, these datasets are costly to obtain in medical contexts and suffer from intra and inter-subject variability. In the present work, two new pre-processing techniques are introduced to improve a deep classifier performance. First, data augmentation based on co-registration is suggested. Then, multi-scale enhancement based on Difference of Gaussians is proposed. Results are accessed in a public mammogram database, the InBreast, in the context of an ordinal problem, the BI-RADS classification. Moreover, a pre-trained Convolutional Neural Network with the AlexNet architecture was used as a base classifier. The multi-class classification experiments show that the proposed pipeline with the Difference of Gaussians and the data augmentation technique outperforms using the original dataset only and using the original dataset augmented by mirroring the images.