Abstract

Abstract

Abstract
We live in a world where data is becoming increasingly valuable and increasingly abundant in volume. All companies produce data from sales, sensors, and various other sources. The main challenges are how can we extract insights from such a rich data environment and if Deep Learning is capable of circumventing Big Data’s challenges. To reach a conclusion, Social Network data is used as a case study for predicting sentiment changes in the Stock Market. The main objective of this paper is to develop a computational study and analyze its performance. Deep Learning was able to handle some challenges of Big Data, allowing results to be obtained and compared with real world situations. The outputs contribute to understand Deep Learning’s usage with Big Data and how it acts in Sentiment Analysis. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Abstract
Fall situations in the context of hospitalization represent high injury risks for the patient and can determine delays or even setbacks in the recovery process. This risk of fall can be minimized by using trained personnel but also using smart equipments that continuously monitor the patients and that can automatically raise alerts as a preventive measure. In this paper a new low-cost system for the prevention of falls in inpatients is proposed. The system is based on a simple modular design that can be easily adapted to distinct equipments such as a bed or an armchair. The sensing devices are pressure mats, transparent for the user and highly sensitive to body position changes, and accelerometers, that can reliably monitor movement while using a very small footprint. The main purpose of the system is to monitor the patients' movement while trying to detect the intention to abandon the bed or armchair. The system was widely tested using a pre-defined protocol and the overall results are quite promising. Some issues were also detected which opens the path for further developments. © 2018 IEEE.

Abstract

Abstract
Even while the scientific community has shown great interest in the analysis of meta-heuristics, the analysis of their parameterization has received little attention. It is the parameterization that will adapt a meta-heuristic to a problem, but it is still performed, mostly, empirically. There are multiple parameterization techniques; however, they are time-consuming, requiring considerable computational effort and they do not take advantage of the meta-heuristics that they parameterize. In order to approach the parameterization of meta-heuristics, in this paper, a self-parameterization framework is proposed. It will automatize the parameterization as an optimization problem, precluding the user from spending too much time on parameterization. The model will automate the parameterization through two meta-heuristics: A meta-heuristic of the solution space and one of the parameter space. To analyze the performance of the framework, a self-parameterization prototype was implemented. The prototype was compared and analyzed in a SP (scheduling problem) and in the TSP (traveling salesman problem). In the SP, the prototype found better solutions than those of the manually parameterized meta-heuristics, although the differences were not statistically significant. In the TSP, the self-parameterization prototype was more effective than the manually parameterized meta-heuristics, this time with statistically significant differences.