Ponencias (Estadística e Investigación Operativa)
URI permanente para esta colecciónhttps://hdl.handle.net/11441/10846
Examinar
Envíos recientes
Ponencia Data-Driven Exploratory Modeling of Temperature and Humidity Effects on Mortality(International Conference on Artificial Intelligence, Computer, Data Sciences and Applications, 2026-02-07) Gómez Losada, Álvaro; Szewczyk, Wojciech; Ciscar, Juan C.; Estadística e Investigación Operativa; FQM153: Estadística e Investigación OperativaThis study investigates the combined influence of temperature and humidity on daily mortality in Brazil and Mexico, using open mortality data and reanalyzed ERA5 climate records from 2000 to 2023. The analysis integrates descriptive statistics, and predictive modeling approaches to evaluate how humidity modulates temperature-related mortality. Results indicate that mortality tends to rise under sub-maximum climatic conditions rather than at extremes, with the highest risks observed for temperature–humidity combinations of 16–36 °C and 9–21 g/kg in Brazil, and 10–35 °C and 3–18 g/kg in Mexico. Optimal lag estimation revealed a three-day delay between exposure and mortality in Brazil, and approximately ten days in Mexico. Random forest and generalized linear models (GLMs) showed limited yet consistent evidence of temperature–humidity interactions affecting mortality.
Ponencia Time series forecasting by recommendation: an empirical analysis on amazon marketplace(Springer, 2019-05-18) Gómez Losada, Álvaro; Duch Brown, Néstor; Estadística e Investigación Operativa; FQM153: Estadística e Investigación OperativaThis study proposes a forecasting methodology for univariate time series (TS) using a Recommender System (RS). The RS is built from a given TS as only input data and following an item-based Collaborative Filtering approach. A set of top-N values is recommended for this TS which represent the forecasts. The idea is to emulate RS elements (the users, items and ratings triple) from the TS. Two TS obtained from Italy’s Amazon webpage were used to evaluate this methodology and very promising performance results were obtained, even the difficult environment chosen to conduct forecasting (short length and unevenly spaced TS). This performance is dependent on the similarity measure used and suffers from the same problems that other RSs (e.g., cold-start). However, this approach does not require high computational power to perform and its intuitive conception allows for being deployed with any programming language.Ponencia Time Series Clustering to Estimate Particulate Matter Contributions from Deserts(MDPI, 2016-07-15) Gómez Losada, Álvaro; Pires, José Carlos M.; Pino Mejías, Rafael; Estadística e Investigación Operativa; FQM153: Estadística e Investigación OperativaExploratory analysis of time series (TS) data is an important approach in experimental studies, with a large range of applications in many different fields, including air pollution studies. To identify structures in single (univariate) TS, main clustering analyses are based on general-purpose clustering algorithms (e.g., k-means, hierarchical clustering methods) and made the assumption that the samples (data) of a TS are independent, ignoring the correlations in consecutive sample values in time. This is specially the case of air pollutant studies based on monitoring data. Air pollutants TS can be studied using TS clustering techniques and as a result, pollution profiles or concentration regimes detected as well as the dependency structure between consecutive data is preserved. Once TS clustering applied over the TS data stream, a set of clusters group the data according to their similar concentration values, and therefore, different pollution profiles can be defined and their estimated range of concentration values. Hidden Markov Models (HMMs) are flexible general-purpose models for univariate and multivariate TS. The TS data are assumed to have a Markov property, and may be viewed as the results of a probabilistic walk along a fixed set of (no directly observable) states. This class of approach considers that each TS is generated by a mixture of underlying probability distributions, typically the Gaussian ones. In this study, HMMs were applied to cluster daily average particulate matter with aerodynamic diameter of 10 μm or less (PM10) TS collected at background monitoring stations from the Iberian Peninsula and Canarian Archipelago (Spain). As a result, PM10 concentration regimes were studied and in particular, the contribution to PM10 ambient concentration levels from the regimes associated to transport of air masses from North Africa deserts was estimated. Regarding this last contribution, we later compared to those obtained using the monthly moving 40th percentile (P40) method over the same TS and no significant quantitative differences were detected. However, the results obtained with HMMs seem to correct the net load of PM10 given by the P40 method, and attributes less impact on areas suffering greater influence from African episodes. The method proposed in this work to estimate PM10 from deserts could improve the P40 method in two ways since it avoids: (i) the smoothed effect which is implicit in the P40 methods after applying a mobile procedure in the TS treatment; and (ii) the empirical approach based on a correlation analysis applied in order to select this particular percentile (40th). Moreover, the use of statistical replicative techniques (bootstrap) together with HMMs has let to obtain an interval confidence in the PM10 contribution estimates from North African deserts. This methodology may be used to estimate particulate matter contributions from any desert; however, a consensus among experts is required to give the regimes obtained with HMMs a definition.Ponencia Competing for Amazon’s Buy Box: A Machine-Learning Approach(Springer, 2019-12-19) Gómez Losada, Álvaro; Duch Brown, Néstor; Estadística e Investigación Operativa; FQM153: Estadística e Investigación OperativaA key feature of the Amazon marketplace is that multiple sellers can sell the same product. In such cases, Amazon recommends one of the sellers to customers in the so-called ‘buy-box’. In this study, the dynamics among sellers for occupying the buy-box was modelled using a classification approach. Italy’s Amazon webpage was crawled during ten months and features from products analyzed to estimate the more relevant ones Amazon could consider for a seller occupy the buy-box. Predictive models showed that the more relevant features are the ratio between consecutive prices in products and their number of assessment received by customers.Artículo A review on differentiability and optimality conditions in fuzzy environments(Springer, 2022-07-04) Hernández Jiménez, Beatriz; Osuna Gómez, Rafaela; Chalco Cano, Yurilev; Costa, Tiago Mendoça da; Estadística e Investigación OperativaIn this paper we present a review of the most important notions and characterizations of differentiability and necessary optimality conditions for a fuzzy multiobjective problem. As basis of this review, we first study the fundamental aspects of the notions of differentiability for interval valued functions, since the fuzzy environment and the interval environment are closely related. Those aspects are related to the different definitions of difference for intervals and their drawbacks, the different definitions and characterizations of the differentiability for interval-valued functions and their drawbacks and how they have been solved in the literature. Based on the most important and meaning results on interval valued functions you can find in the literature, a review on notions of differentiability in fuzzy context is given, both in the case of functions of one variable, and several variables. And finally we present the review results of the necessary optimality conditions for fuzzy multiobjective problems and the main conclusions.Ponencia Optimisation of Aiming Strategies in Solar Tower Power Plants(AIP, 2018-11-08) Ashley, Thomas Ian; Carrizosa Priego, Emilio José; Fernández Cara, Enrique; Estadística e Investigación Operativa; FQM329: OptimizacionInclement weather effects have a direct impact on the efficiency of a Solar Power Tower plant and have the potential to damage the receiver by flash heating. An optimised aiming strategy for the heliostat field mitigates the risk of receiver damage and maximises plant efficiency. A stochastic integer programming approach is applied to optimise the aiming strategy of the heliostat field, with uncertainty in the cloud location, size and density. The optimisation technique is demonstrated with a test case and results are presented for near real-time simulation of the optimal aiming strategy.Ponencia Convex analysis applied to location theory(2007) Puerto Albandoz, Justo; Rodríguez Chía, Antonio Manuel; Estadística e Investigación Operativa; FQM331: Métodos y Modelos de la Estadística y la Investigación OperativaPonencia Population pharmacokinetics of colistin: implications for clinical use for Gram-negative pathogens(2016) Pachón Ibáñez, María Eugenia; Docobo Pérez, Fernando; Blanquero Bravo, Rafael; Garnacho Montero, José; Benítez Peña, Sandra; Dimopoulos, George; Rosso Fernández, Clara; Gutiérrez Pizarraya, Antonio; López Cortés, Luis Fernando; Cisneros Herreros, José Miguel; Estadística e Investigación Operativa; FQM329: Optimización; CTS210: Adherencia Bacteriana a Nuevos BiomaterialesThe objective of this study was to characterize the pharmacokinetics of colistin methanesulphonate (CMS) and colistin in critically ill patients following the administration of a 4.5 MU CMS loading dose follow by 3MU CMS Q8. A population PK model and Monte Carlo simulation were used to calculate the probability of target attainment (PTA) against Acinetobacter baumannii and Pseudomonas aeruginosa by considering a range of MIC values seen in the clinic.Ponencia Ordered weighted average optimization in multiobjective spanning tree problems(2015) Fernández Areizaga, Elena; Pozo Montaño, Miguel Ángel; Puerto Albandoz, Justo; Scozzari, Andrea; Estadística e Investigación Operativa; FQM241: Grupo de Investigación en Localización; FQM331: Métodos y Modelos de la Estadística y la Investigación OperativaPonencia Multisource linear regression(2015) Blanco Izquierdo, Víctor; Ponce López, Diego; Puerto Albandoz, Justo; Estadística e Investigación Operativa; FQM331: Métodos y Modelos de la Estadística y la Investigación OperativaPonencia Robust p-median problem with vector autoregressive demands(2015) Carrizosa Priego, Emilio José; Olivares Nadal, Alba Victoria; Ramírez Cobo, Josefa; Estadística e Investigación Operativa; FQM329: OptimizaciónPonencia Un esquema general de búsqueda local en programación entera. Evaluación computacional(1995) Mayor Gallego, José Antonio; Ruiz Canales, Pascual; Estadística e Investigación Operativa; FQM331: Métodos y Modelos de la Estadística y la Investigación Operativa; FQM153: Estadística e Investigación OperativaEn este trabajo, estudiamos un procedimiento general de búsqueda local basado en un sistema probabilístico de entornos, que puede ser combinada con técnicas de “annealing” simulado y de banda inferior. Para ello, se estudia un sistema de entornos definidos sobre conjuntos discretos y que incorporan información proporcionada por el gradiente de la función objetivo. Adicionalmente, se estudia el comportamiento de este tipo de entornos en combinación con técnicas de “simulated annealing” en lo que respecta a la convergencia asintótica a óptimos globales. Finalmente se exponen una serie de resultados computacionales obtenidos al aplicar estas técnicas a varios problemas de programación entera, tanto lineal como no lineal.Ponencia Generalizations of nondifferentiable convex functions and some characterizations(1991) Ruiz Canales, Pascual; Beato Moreno, Antonio; Estadística e Investigación Operativa; FQM153: Estadística e Investigación Operativa; FQM331: Métodos y Modelos de la Estadística y la Investigación OperativaIn this paper we generalize the convex functions, defining the concept of preconvex function and we study some characterizations by intervals, some characterizations by polytopes, some characterizations by level sets, some properties of the extreme points and some relations whith the convex functions. Also, we define the R-quasiconvex functions as a generalization of the quasiconvex functions, and we study some characterizations by level sets and by separation sets, and some relations with the quasiconvex functions.Ponencia Estimación del índice de Gini mediante diseños muestrales complejos(2007) Mayor Gallego, José Antonio; Estadística e Investigación Operativa; FQM331: Métodos y Modelos de la Estadística y la Investigación OperativaPonencia Algunos modelos de localizacion de un servicio bajo un escenario de demanda incierto(2007) Conde Sánchez, Eduardo; Estadística e Investigación Operativa; FQM331: Métodos y Modelos de la Estadística y la Investigación OperativaPonencia Modelo estadístico para la predicción del índice estandarizado de sequía pluviométrica (IESP) en Andalucía(Universidad de Salamanca, 2012) Blanquero Bravo, Rafael; Carrizosa Priego, Emilio José; Pita López, María Fernanda; Camarillo Naranjo, Juan Mariano; Álvarez Francoso, José Ignacio; Estadística e Investigación Operativa; Geografía Física y Análisis Geográfico RegionalLa comunicación aborda el diseño de un modelo estadístico de predicción dinámica de la sequía en Andalucía y su persistencia en un horizonte temporal de 12 meses a partir de los datos históricos (1950-2012) del Índice Estandarizado de Sequía Pluviométrica (IESP) en 243 observatorios de Andalucía. Se emplea un algoritmo kNN (k-Nearest Neighbors), que busca las situaciones pasadas más similares a la actual y predice el futuro promediando lo que ocurrió a continuación en dichas situaciones. Los resultados producen porcentajes de error muy reducidos. El modelo se está aplicando en rutina en el Sistema de Información de Climatología Ambiental (CLIMA) de la Consejería de Medio Ambiente de la Junta de Andalucía (http://www.climasig.es).