Examinando por Autor "Alfaro, Cesar"

Mostrando 1 - 6 de 6

Adapting support vector optimisation algorithms to textual gender classification
(Springer, 2024-04-13) Gómez, Javier; Alfaro, Cesar; Ortega, Felipe; Moguerza, Javier M.; Algar, Maria Jesus; Moreno, Raul
In this paper, we focus on the problem of determining the gender of the person described in a biographical text. Since support vector machine classifiers are well suited for text classification tasks, we present a new stopping criterion for support vector optimisation algorithms tailored to this problem. This new approach exploits the geometric properties of the vector representation of such content. An experiment on a set of English and Spanish biographical articles retrieved from Wikipedia illustrates this approach and compares it to other machine learning classification algorithms. The proposed method allows real-time classification algorithm training. Moreover, these results confirm the advantage of leveraging additional gender information in strongly inflected languages, like Spanish, for this task
ESPRES: A web application for interactive analysis of multiple pressures in aquatic ecosystems
(Elsevier, 2020-07-11) Udias, Angel; Pistocchi, Alberto; Vigiak, Olga; Grizzetti, Bruna; Bouraoui, Faycal; Alfaro, Cesar
ESPRES (Efficient Strategies for anthropogenic Pressure Reduction in European waterSheds) is a web-based Decision Support System (DSS) designed to explore management options for achieving environmental targets in European freshwaters. The tool integrates multi-objective optimization (MOO) algorithms for selecting the best management options in a river basin and models assessing the consequent changes in the water quantity (water flow) and quality (nutrient concentration). The MOO engine identifies Pareto front strategies that are trade-offs between environmental objectives for water bodies and the effort required for reducing the pressures. The web interface provides tools to set the effort perceived by different river basin stakeholders considering technical feasibility, political difficulty, and social acceptability of the alternative options. The environmental impact of management options (scenarios) is assessed with models developed at the European scale. ESPRES enables comparison of management solutions and allows quantifying environmental and socio-economic trade-offs inherent to the decision making process.
ESPRES: A web application for interactive analysis of multiple pressures in aquatic ecosystems
(Elsevier, 2020-11-20) Udias, Angel; Pistocchi, Alberto; Vigiak, Olga; Grizzetti, Bruna; Bouraoui, Faycal; Alfaro, Cesar
ESPRES (Efficient Strategies for anthropogenic Pressure Reduction in European waterSheds) is a web-based Decision Support System (DSS) designed to explore management options for achieving environmental targets in European freshwaters. The tool integrates multi-objective optimization (MOO) algorithms for selecting the best management options in a river basin and models assessing the consequent changes in the water quantity (water flow) and quality (nutrient concentration). The MOO engine identifies Pareto front strategies that are trade-offs between environmental objectives for water bodies and the effort required for reducing the pressures. The web interface provides tools to set the effort perceived by different river basin stakeholders considering technical feasibility, political difficulty, and social acceptability of the alternative options. The environmental impact of management options (scenarios) is assessed with models developed at the European scale. ESPRES enables comparison of management solutions and allows quantifying environmental and socio-economic trade-offs inherent to the decision making process.
Forecasting and assessing consequences of aviation safety occurrences
(Elsevier, 2019-01) Rios Insua, David; Alfaro, Cesar; Gomez, Javier; Hernandez-Coronado, Pablo; Bernal, Francisco
Aviation safety is essential for the healthy growth and sustainability of the global economy. The implementation of Safety Management Systems to support safe service delivery has become one of the most important goals within the airline industry over the last years. However, in most cases the involved organisations use unsophisticated methods based on risk matrices for the development of such systems. In this paper, we present models to forecast and assess the consequences of aviation safety occurrences as part of a framework for aviation safety risk management at state level.
Forecasting aviation safety occurrences
(John Wiley & Sons Ltd., 2022-06-17) Flores, Bruno; Rios Insua, David; Alfaro, Cesar; Gomez, Javier
We present a general framework for aviation safety occurrence forecasting. This is a major component of a methodology for aviation safety risk management at national level. It covers novel models as well as novel combinations of earlier models. Having good quality occurrence and severity forecasting models is paramount to properly manage risks, maintain the confidence of its users and preserve the status of aviation as a safe transportation mode. The problem is involved due to the presence of complex effects like seasonality, trends, or stress that impact the rates of various occurrences and the uncertainty about future number of operations.
Toward Accelerated Training of Parallel Support Vector Machines Based on Voronoi Diagrams
(MDPI, 2021-11-29) Alfaro, Cesar; Gomez, Javier; M. Moguerza, Javier; Castillo, Javier; Martinez, Jose I.
Typical applications of wireless sensor networks (WSN), such as in Industry 4.0 and smart cities, involves acquiring and processing large amounts of data in federated systems. Important challenges arise for machine learning algorithms in this scenario, such as reducing energy consumption and minimizing data exchange between devices in different zones. This paper introduces a novel method for accelerated training of parallel Support Vector Machines (pSVMs), based on ensembles, tailored to these kinds of problems. To achieve this, the training set is split into several Voronoi regions. These regions are small enough to permit faster parallel training of SVMs, reducing computational payload. Results from experiments comparing the proposed method with a single SVM and a standard ensemble of SVMs demonstrate that this approach can provide comparable performance while limiting the number of regions required to solve classification tasks. These advantages facilitate the development of energy-efficient policies in WSN.