El presente estudio tiene como objetivo identificar los Estilos de Aprendizaje con que cuentan l... more El presente estudio tiene como objetivo identificar los Estilos de Aprendizaje con que cuentan los alumnos próximos a egresar de dos modalidades de bachillerato (General y Telebachillerato) en las ciudades de Xalapa y Coatepec, Veracruz, México. Para ello se utilizó el Inventario de Estilos de Aprendizaje de Kolb (IEA) en una muestra de 162 estudiantes del último semestre en las diferentes áreas terminales (Químico-Biológico, Físico-Matemático, Económico-Administrativo y Humanidades-Ciencias Sociales). Los resultados obtenidos muestran que el estilo Asimilador predominó en el bachillerato general, en tanto que para el telebachillerato predominaron el estilo Asimilador y el Divergente. Estadísticamente se encontró que los Estilos de Aprendizaje caracterizados no tienen relación alguna (p>0.05) con el sexo de los estudiantes, el tipo de bachillerato o el área terminal que cursan. Así mismo la interacción de dichas variables tampoco tiene un efecto sobre la presencia de estos estil...
El uso diario de las imágenes médicas en los traumatismos de miembro inferior y su prevalencia en... more El uso diario de las imágenes médicas en los traumatismos de miembro inferior y su prevalencia en el segmento tobillo-pie, fue lo que estimuló la realización de este trabajo. Debido a ello decidimos correlacionar cortes cadavéricos de tipo coronal, axial y sagital con sus radiografías, para brindar una guía al enfoque diagnóstico del tobillo-pie. Como resultado obtuvimos la disposición particular que adquieren en el espacio, las estructuras anatómicas según la proyección del corte, con sus interrelaciones profundas y de superficie; ajustándonos al nuevo paradigma: la anatomía espacial en imágenes diagnósticas de tomografía axial computada y resonancia magnética. La observación metódica y sistemática de elementos del tobillo-pie en cortes, es una contribución pedagógica, aplicable a la interpretación topográfica de imágenes normales, que serviría de ayuda al estudiante y médico general.
arXiv: Neural and Evolutionary Computing, Jul 24, 2015
Restricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain gener... more Restricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain generative models of data distributions. RBMs are often trained using the Contrastive Divergence learning algorithm (CD), an approximation to the gradient of the data log-likelihood. A simple reconstruction error is often used as a stopping criterion for CD, although several authors [1], [2] have raised doubts concerning the feasibility of this procedure. In many cases the evolution curve of the reconstruction error is monotonic while the log-likelihood is not, thus indicating that the former is not a good estimator of the optimal stopping point for learning. However, not many alternatives to the reconstruction error have been discussed in the literature. In this manuscript we investigate simple alternatives to the reconstruction error, based on the inclusion of information contained in neighboring states to the training set, as a stopping criterion for CD learning.
Probabilistic models based on Restricted Boltzmann Machines (RBMs) imply the evaluation of normal... more Probabilistic models based on Restricted Boltzmann Machines (RBMs) imply the evaluation of normalized Boltzmann factors, which in turn require from the evaluation of the partition function Z. The exact evaluation of Z, though, becomes a forbiddingly expensive task as the system size increases. This even worsens when one considers most usual learning algorithms for RBMs, where the exact evaluation of the gradient of the log-likelihood of the empirical distribution of the data includes the computation of Z at each iteration. The Annealed Importance Sampling (AIS) method provides a tool to stochastically estimate the partition function of the system. So far, the standard use of the AIS algorithm in the Machine Learning context has been done using a large number of Monte Carlo steps. In this work we show that this may not be required if a proper starting probability distribution is employed as the initialization of the AIS algorithm. We analyze the performance of AIS in both small- and ...
2020 International Joint Conference on Neural Networks (IJCNN), 2020
Machine learning (ML) methods have shown great potential for the analysis of data involved in med... more Machine learning (ML) methods have shown great potential for the analysis of data involved in medical decisions. However, for these methods to be incorpored in the medical pipeline, they must be made interpretable not only to the data analyst, but also to the medical expert. In this work, we have applied a combination of feature transformation, selection and classification using ML and statistical methods to differentiate between control (untreated) and Temozolomide (TMZ)-treated tumour tissue from a glioblastoma (brain tumour) murine model. As input, we have used T2 weighted magnetic resonance images (MRI) and spectroscopic imaging (MRSI). Radiomics features have been extracted from the MRI dataset, while convex Non-negative Matrix Factorization (Convex-NMF) was used to extract sources from the MRSI dataset. Exhaustive feature selection has revealed parsimonious feature subsets that facilitate the expert interpretation of results while retaining a high discriminatory ability.
Electronic Government, An International Journal, 2021
The tropical cyclones on different stages depression, storm and hurricane, have been topics of re... more The tropical cyclones on different stages depression, storm and hurricane, have been topics of research in recent decades. This work studies the spatial variability of the precipitation generated by the tropical storm Cristobal, which, since its formation and passing through the Yucatan Peninsula, Mexico. This tropical disturbance arose in late May from tropical depression number three. Cristobal was declared as a tropical storm in the Campeche Sonda, Mexico on May 31 before the beginning of the official Atlantic season. Cristobal stayed abnormally for three days in the southern Gulf of Mexico, entering the Campeche entity to move north, and ending up affecting Louisiana, United States as a tropical storm. In the Yucatan Peninsula, daily rainfall exceeded over 240% of June average (110-220 mm). The historical record in four meteorological stations and the accumulated rainfall registered in six days represent 90% of the annual average in certain areas, which damaged agricultural acti...
With the purpose to demonstrate risk factors associated to parasites and iron depletion on school... more With the purpose to demonstrate risk factors associated to parasites and iron depletion on school children it was carried out a case and control design, 102 children of 60 to 144 months of age as a...
The results of a comprehensive experimental program focused on the study of progressive degradati... more The results of a comprehensive experimental program focused on the study of progressive degradation processes on a low permeability argillaceous rock induced by hydraulic cycles is presented in the paper. Relative humidity cycles were applied with vapour transfer technique, and bender elements were used to evaluate the evolution of shear stiffness during the application of hydraulic paths. The characterisation of the material included the determination of the water retention properties (water retention curve), as well as the pore size distribution using mercury intrusion porosimetry, complemented with SEM micrographs and elemental analysis using X-ray spectroscopy to characterise micro-structural features and detect clayey and non-clayey constituents. Results showed cumulative and irreversible swelling of the samples with the application of relative humidity cycles. Regarding shear stiffness, an important reduction (around 80%) was observed at the end of the hydraulic paths applied....
EN LA INAUGURACION DEL CURSO DE POSGRADO DE ENFERMERIA DEL ADULTO EN ESTADO CRITICO, QUE IMPARTEN... more EN LA INAUGURACION DEL CURSO DE POSGRADO DE ENFERMERIA DEL ADULTO EN ESTADO CRITICO, QUE IMPARTEN LA UNAM Y LA UNIDAD MEDICA DE ALTA ESPECIALIDAD (UMAE) DEL HOSPITAL DE ESPECIALIDADES CENTRO MEDICO NACIONAL SIGLO XXI, SEVERINO RUBIO DOMINGUEZ, DIRECTOR DE LA ESCUELA NACIONAL DE ENFERMERIA Y OBSTETRICIA (ENEO), SENALO QUE LA ENFERMERIA ES CADA VEZ MAS COMPLEJA; SU ACTIVIDAD SE HA ESPECIALIZADO MAS EN LO ACADEMICO Y LO TECNOLOGICO. SU EJERCICIO PROFESIONAL ESTA ENFOCADO AL CUIDADO DE LA VIDA Y MAS EN CIRCUNSTANCIAS CRITICAS EN LAS QUE LA PERSONA CORRE RIESGO DE MUERTE A CAUSA DE UNA ENFERMEDAD; DE AHI LA IMPORTANCIA DEL CURSO, SUBRAYO. LA ESPECIALIZACION ES RESULTADO DE UN CONVENIO GENERAL DE COLABORACION FIRMADO HACE UNOS ANOS POR AMBAS INSTITUCIONES PARA FORMAR, CAPACITAR Y COORDINAR LA PARTE DE EDUCACION E INVESTIGACION DEL PERSONAL DE SALUD. ANTE FUNCIONARIOS Y autorIDADES DE LA UNAM Y DEL IMSS, PROFESORES Y ALUMNOS DEL CURSO, SEVERINO RUBIO APUNTO QUE ESTE PROGRAMA CONJUNTO DE FO...
The increasing access to brain signal data using electroencephalography creates new opportunities... more The increasing access to brain signal data using electroencephalography creates new opportunities to study electrophysiological brain activity and perform ambulatory diagnoses of neurological disorders. This work proposes a pairwise distance learning approach for Schizophrenia classification relying on the spectral properties of the signal. Given the limited number of observations (i.e. the case and/or control individuals) in clinical trials, we propose a Siamese neural network architecture to learn a discriminative feature space from pairwise combinations of observations per channel. In this way, the multivariate order of the signal is used as a form of data augmentation, further supporting the network generalization ability. Convolutional layers with parameters learned under a cosine contrastive loss are proposed to adequately explore spectral images derived from the brain signal. Results on a case-control population show that the features extracted using the proposed neural network lead to an improved Schizophrenia diagnosis (+10pp in accuracy and sensitivity) against baselines, suggesting the existence of nontrivial electrophysiological brain patterns able to capture discriminative neuroplasticity profiles among individuals.
Learning algorithms for energy based Boltzmann architectures that rely on gradient descent are in... more Learning algorithms for energy based Boltzmann architectures that rely on gradient descent are in general computationally prohibitive, typically due to the exponential number of terms involved in computing the partition function. In this way one has to resort to approximation schemes for the evaluation of the gradient. This is the case of Restricted Boltzmann Machines (RBM) and its learning algorithm Contrastive Divergence (CD). It is well-known that CD has a number of shortcomings, and its approximation to the gradient has several drawbacks. Overcoming these defects has been the basis of much research and new algorithms have been devised, such as persistent CD. In this manuscript we propose a new algorithm that we call Weighted CD (WCD), built from small modifications of the negative phase in standard CD. However small these modifications may be, experimental work reported in this paper suggest that WCD provides a significant improvement over standard CD and persistent CD at a small additional computational cost.
Biocuration in the omics sciences has become paramount, as research in these fields rapidly evolv... more Biocuration in the omics sciences has become paramount, as research in these fields rapidly evolves towards increasingly data-dependent models. As a result, the management of web-accessible publicly-available databases becomes a central task in biological knowledge dissemination. One relevant challenge for biocurators is the unambiguous identification of biological entities. In this study, we illustrate the adequacy of machine learning methods as biocuration assistance tools using a publicly available protein database as an example. This database contains information on G Protein-Coupled Receptors (GPCRs), which are part of eukaryotic cell membranes and relevant in cell communication as well as major drug targets in pharmacology. These receptors are characterized according to subtype labels. Previous analysis of this database provided evidence that some of the receptor sequences could be affected by a case of label noise, as they appeared to be too consistently misclassified by mach...
Estamos asistiendo, en los últimos años, a un debate académico sobre la conveniencia y el papel q... more Estamos asistiendo, en los últimos años, a un debate académico sobre la conveniencia y el papel que debería tener la evaluación del ciudadano en los sistemas de medición del desempeño o rendimiento (performance) de la Administración. Después de considerar las reservas de algunos autores, nos decantamos por la literatura que acentúa el carácter estratégico de la opinión del ciudadano. El artículo explora la viabilidad de una nueva escala para evaluar la calidad, percibida por el ciudadano, de los servicios de la Administración local. No hemos utilizado la escala SERVQUAL, por considerarla menos adecuada por varias razones, entre las que destaca la «incompatibilidad fáctica» de su aplicación. En su lugar, proponemos la escala multidimensional jerárquica como alternativa con más posibilidades. A partir de los datos de una encuesta realizada en varios municipios del Levante español, se ha procedido a realizar la validación empírica de la nueva escala, a fin de comprobar su capacidad predictiva sobre la intención favorable del ciudadano hacia la Administración local. Palabras clave: medición del desempeño; administración local; escala SERVQUAL; escala multidimensional jerárquica; calidad percibida; evaluación ciudadana; modelos de ecuaciones estructurales. 1. Este trabajo ha sido posible, en parte, gracias a sendos proyectos de I+D+i (del MEC y del MICINN: SEJ2004-01098 y CSO2008-03337), dirigidos ambos por Benjamín González Rodríguez, coautor de este artículo.
Cada vez son más los artículos que consideran la opinión ciudadana como información esencial para... more Cada vez son más los artículos que consideran la opinión ciudadana como información esencial para evaluar la acción administrativa local. La investigación reciente viene demostrando la presencia de un vínculo empírico entre la calidad percibida de los servicios locales y la intención de conducta favorable al titular del gobierno local. Debido a que la literatura científica también acumula evidencia de relación entre ideología y preferencia política, el propósito de este artículo ha consistido en averiguar cómo interviene la ideología en el impacto de la calidad percibida sobre la preferencia política, es decir, si lo anula o lo modera y en qué grado. Como los conceptos calidad percibida e intención de conducta son constructos enlazados causalmente, el tratamiento ha requerido el empleo de la metodología de ecuaciones estructurales (SEM). El análisis progresa verificando las hipótesis de espuridad y moderación, pero los datos van a rechazarlas y apuntar a un mecanismo explicativo no esperado. Una encuesta realizada en nueve localidades de la provincia de Valencia (España) ha permitido comprobar el sorprendente papel que la ideología está desempeñando en la formación de la intención política del ciudadano a partir de las evaluaciones del servicio. Palabras clave: evaluación de los servicios; escala multidimensional jerárquica; calidad percibida; opinión pública; modelos de ecuaciones estructurales; análisis de invarianza; ideología política; Administración local; preferencia política.
Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005.
An experimental study of several decision issues for wrapper Feature Selection with Multi-Layer P... more An experimental study of several decision issues for wrapper Feature Selection with Multi-Layer Perceptrons is presented, namely the stopping criterion, the data set where the saliency is measured and the network retraining before computing the saliency. Experimental results with the Sequential Backward Selection procedure indicate that the increase in the computational cost associated with retraining the network with every feature temporarily removed before computing the saliency is rewarded with a significant performance improvement. Despite being quite intuitive, this idea has been hardly used in practice. Regarding the stopping criterion and the data set where the saliency is measured, the procedure profits from measuring the saliency in a validation set, as reasonably expected. A somehow non-intuitive conclusion can be drawn by looking at the stopping criterion, where it is suggested that forcing overtraining may be as useful as early stopping. 1 There is no commonly accepted definition of the relevance of a variable (see [3], [7], for example). Given a data set, we consider that a variable is irrelevant for a Supervised Machine Learning system when its optimal performance is not affected negatively by the absence of that variable ([7], page 29). Note that this is a dynamic definition, since the relevance of a variable may be affected by the presence or absence of other ones.
El presente estudio tiene como objetivo identificar los Estilos de Aprendizaje con que cuentan l... more El presente estudio tiene como objetivo identificar los Estilos de Aprendizaje con que cuentan los alumnos próximos a egresar de dos modalidades de bachillerato (General y Telebachillerato) en las ciudades de Xalapa y Coatepec, Veracruz, México. Para ello se utilizó el Inventario de Estilos de Aprendizaje de Kolb (IEA) en una muestra de 162 estudiantes del último semestre en las diferentes áreas terminales (Químico-Biológico, Físico-Matemático, Económico-Administrativo y Humanidades-Ciencias Sociales). Los resultados obtenidos muestran que el estilo Asimilador predominó en el bachillerato general, en tanto que para el telebachillerato predominaron el estilo Asimilador y el Divergente. Estadísticamente se encontró que los Estilos de Aprendizaje caracterizados no tienen relación alguna (p>0.05) con el sexo de los estudiantes, el tipo de bachillerato o el área terminal que cursan. Así mismo la interacción de dichas variables tampoco tiene un efecto sobre la presencia de estos estil...
El uso diario de las imágenes médicas en los traumatismos de miembro inferior y su prevalencia en... more El uso diario de las imágenes médicas en los traumatismos de miembro inferior y su prevalencia en el segmento tobillo-pie, fue lo que estimuló la realización de este trabajo. Debido a ello decidimos correlacionar cortes cadavéricos de tipo coronal, axial y sagital con sus radiografías, para brindar una guía al enfoque diagnóstico del tobillo-pie. Como resultado obtuvimos la disposición particular que adquieren en el espacio, las estructuras anatómicas según la proyección del corte, con sus interrelaciones profundas y de superficie; ajustándonos al nuevo paradigma: la anatomía espacial en imágenes diagnósticas de tomografía axial computada y resonancia magnética. La observación metódica y sistemática de elementos del tobillo-pie en cortes, es una contribución pedagógica, aplicable a la interpretación topográfica de imágenes normales, que serviría de ayuda al estudiante y médico general.
arXiv: Neural and Evolutionary Computing, Jul 24, 2015
Restricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain gener... more Restricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain generative models of data distributions. RBMs are often trained using the Contrastive Divergence learning algorithm (CD), an approximation to the gradient of the data log-likelihood. A simple reconstruction error is often used as a stopping criterion for CD, although several authors [1], [2] have raised doubts concerning the feasibility of this procedure. In many cases the evolution curve of the reconstruction error is monotonic while the log-likelihood is not, thus indicating that the former is not a good estimator of the optimal stopping point for learning. However, not many alternatives to the reconstruction error have been discussed in the literature. In this manuscript we investigate simple alternatives to the reconstruction error, based on the inclusion of information contained in neighboring states to the training set, as a stopping criterion for CD learning.
Probabilistic models based on Restricted Boltzmann Machines (RBMs) imply the evaluation of normal... more Probabilistic models based on Restricted Boltzmann Machines (RBMs) imply the evaluation of normalized Boltzmann factors, which in turn require from the evaluation of the partition function Z. The exact evaluation of Z, though, becomes a forbiddingly expensive task as the system size increases. This even worsens when one considers most usual learning algorithms for RBMs, where the exact evaluation of the gradient of the log-likelihood of the empirical distribution of the data includes the computation of Z at each iteration. The Annealed Importance Sampling (AIS) method provides a tool to stochastically estimate the partition function of the system. So far, the standard use of the AIS algorithm in the Machine Learning context has been done using a large number of Monte Carlo steps. In this work we show that this may not be required if a proper starting probability distribution is employed as the initialization of the AIS algorithm. We analyze the performance of AIS in both small- and ...
2020 International Joint Conference on Neural Networks (IJCNN), 2020
Machine learning (ML) methods have shown great potential for the analysis of data involved in med... more Machine learning (ML) methods have shown great potential for the analysis of data involved in medical decisions. However, for these methods to be incorpored in the medical pipeline, they must be made interpretable not only to the data analyst, but also to the medical expert. In this work, we have applied a combination of feature transformation, selection and classification using ML and statistical methods to differentiate between control (untreated) and Temozolomide (TMZ)-treated tumour tissue from a glioblastoma (brain tumour) murine model. As input, we have used T2 weighted magnetic resonance images (MRI) and spectroscopic imaging (MRSI). Radiomics features have been extracted from the MRI dataset, while convex Non-negative Matrix Factorization (Convex-NMF) was used to extract sources from the MRSI dataset. Exhaustive feature selection has revealed parsimonious feature subsets that facilitate the expert interpretation of results while retaining a high discriminatory ability.
Electronic Government, An International Journal, 2021
The tropical cyclones on different stages depression, storm and hurricane, have been topics of re... more The tropical cyclones on different stages depression, storm and hurricane, have been topics of research in recent decades. This work studies the spatial variability of the precipitation generated by the tropical storm Cristobal, which, since its formation and passing through the Yucatan Peninsula, Mexico. This tropical disturbance arose in late May from tropical depression number three. Cristobal was declared as a tropical storm in the Campeche Sonda, Mexico on May 31 before the beginning of the official Atlantic season. Cristobal stayed abnormally for three days in the southern Gulf of Mexico, entering the Campeche entity to move north, and ending up affecting Louisiana, United States as a tropical storm. In the Yucatan Peninsula, daily rainfall exceeded over 240% of June average (110-220 mm). The historical record in four meteorological stations and the accumulated rainfall registered in six days represent 90% of the annual average in certain areas, which damaged agricultural acti...
With the purpose to demonstrate risk factors associated to parasites and iron depletion on school... more With the purpose to demonstrate risk factors associated to parasites and iron depletion on school children it was carried out a case and control design, 102 children of 60 to 144 months of age as a...
The results of a comprehensive experimental program focused on the study of progressive degradati... more The results of a comprehensive experimental program focused on the study of progressive degradation processes on a low permeability argillaceous rock induced by hydraulic cycles is presented in the paper. Relative humidity cycles were applied with vapour transfer technique, and bender elements were used to evaluate the evolution of shear stiffness during the application of hydraulic paths. The characterisation of the material included the determination of the water retention properties (water retention curve), as well as the pore size distribution using mercury intrusion porosimetry, complemented with SEM micrographs and elemental analysis using X-ray spectroscopy to characterise micro-structural features and detect clayey and non-clayey constituents. Results showed cumulative and irreversible swelling of the samples with the application of relative humidity cycles. Regarding shear stiffness, an important reduction (around 80%) was observed at the end of the hydraulic paths applied....
EN LA INAUGURACION DEL CURSO DE POSGRADO DE ENFERMERIA DEL ADULTO EN ESTADO CRITICO, QUE IMPARTEN... more EN LA INAUGURACION DEL CURSO DE POSGRADO DE ENFERMERIA DEL ADULTO EN ESTADO CRITICO, QUE IMPARTEN LA UNAM Y LA UNIDAD MEDICA DE ALTA ESPECIALIDAD (UMAE) DEL HOSPITAL DE ESPECIALIDADES CENTRO MEDICO NACIONAL SIGLO XXI, SEVERINO RUBIO DOMINGUEZ, DIRECTOR DE LA ESCUELA NACIONAL DE ENFERMERIA Y OBSTETRICIA (ENEO), SENALO QUE LA ENFERMERIA ES CADA VEZ MAS COMPLEJA; SU ACTIVIDAD SE HA ESPECIALIZADO MAS EN LO ACADEMICO Y LO TECNOLOGICO. SU EJERCICIO PROFESIONAL ESTA ENFOCADO AL CUIDADO DE LA VIDA Y MAS EN CIRCUNSTANCIAS CRITICAS EN LAS QUE LA PERSONA CORRE RIESGO DE MUERTE A CAUSA DE UNA ENFERMEDAD; DE AHI LA IMPORTANCIA DEL CURSO, SUBRAYO. LA ESPECIALIZACION ES RESULTADO DE UN CONVENIO GENERAL DE COLABORACION FIRMADO HACE UNOS ANOS POR AMBAS INSTITUCIONES PARA FORMAR, CAPACITAR Y COORDINAR LA PARTE DE EDUCACION E INVESTIGACION DEL PERSONAL DE SALUD. ANTE FUNCIONARIOS Y autorIDADES DE LA UNAM Y DEL IMSS, PROFESORES Y ALUMNOS DEL CURSO, SEVERINO RUBIO APUNTO QUE ESTE PROGRAMA CONJUNTO DE FO...
The increasing access to brain signal data using electroencephalography creates new opportunities... more The increasing access to brain signal data using electroencephalography creates new opportunities to study electrophysiological brain activity and perform ambulatory diagnoses of neurological disorders. This work proposes a pairwise distance learning approach for Schizophrenia classification relying on the spectral properties of the signal. Given the limited number of observations (i.e. the case and/or control individuals) in clinical trials, we propose a Siamese neural network architecture to learn a discriminative feature space from pairwise combinations of observations per channel. In this way, the multivariate order of the signal is used as a form of data augmentation, further supporting the network generalization ability. Convolutional layers with parameters learned under a cosine contrastive loss are proposed to adequately explore spectral images derived from the brain signal. Results on a case-control population show that the features extracted using the proposed neural network lead to an improved Schizophrenia diagnosis (+10pp in accuracy and sensitivity) against baselines, suggesting the existence of nontrivial electrophysiological brain patterns able to capture discriminative neuroplasticity profiles among individuals.
Learning algorithms for energy based Boltzmann architectures that rely on gradient descent are in... more Learning algorithms for energy based Boltzmann architectures that rely on gradient descent are in general computationally prohibitive, typically due to the exponential number of terms involved in computing the partition function. In this way one has to resort to approximation schemes for the evaluation of the gradient. This is the case of Restricted Boltzmann Machines (RBM) and its learning algorithm Contrastive Divergence (CD). It is well-known that CD has a number of shortcomings, and its approximation to the gradient has several drawbacks. Overcoming these defects has been the basis of much research and new algorithms have been devised, such as persistent CD. In this manuscript we propose a new algorithm that we call Weighted CD (WCD), built from small modifications of the negative phase in standard CD. However small these modifications may be, experimental work reported in this paper suggest that WCD provides a significant improvement over standard CD and persistent CD at a small additional computational cost.
Biocuration in the omics sciences has become paramount, as research in these fields rapidly evolv... more Biocuration in the omics sciences has become paramount, as research in these fields rapidly evolves towards increasingly data-dependent models. As a result, the management of web-accessible publicly-available databases becomes a central task in biological knowledge dissemination. One relevant challenge for biocurators is the unambiguous identification of biological entities. In this study, we illustrate the adequacy of machine learning methods as biocuration assistance tools using a publicly available protein database as an example. This database contains information on G Protein-Coupled Receptors (GPCRs), which are part of eukaryotic cell membranes and relevant in cell communication as well as major drug targets in pharmacology. These receptors are characterized according to subtype labels. Previous analysis of this database provided evidence that some of the receptor sequences could be affected by a case of label noise, as they appeared to be too consistently misclassified by mach...
Estamos asistiendo, en los últimos años, a un debate académico sobre la conveniencia y el papel q... more Estamos asistiendo, en los últimos años, a un debate académico sobre la conveniencia y el papel que debería tener la evaluación del ciudadano en los sistemas de medición del desempeño o rendimiento (performance) de la Administración. Después de considerar las reservas de algunos autores, nos decantamos por la literatura que acentúa el carácter estratégico de la opinión del ciudadano. El artículo explora la viabilidad de una nueva escala para evaluar la calidad, percibida por el ciudadano, de los servicios de la Administración local. No hemos utilizado la escala SERVQUAL, por considerarla menos adecuada por varias razones, entre las que destaca la «incompatibilidad fáctica» de su aplicación. En su lugar, proponemos la escala multidimensional jerárquica como alternativa con más posibilidades. A partir de los datos de una encuesta realizada en varios municipios del Levante español, se ha procedido a realizar la validación empírica de la nueva escala, a fin de comprobar su capacidad predictiva sobre la intención favorable del ciudadano hacia la Administración local. Palabras clave: medición del desempeño; administración local; escala SERVQUAL; escala multidimensional jerárquica; calidad percibida; evaluación ciudadana; modelos de ecuaciones estructurales. 1. Este trabajo ha sido posible, en parte, gracias a sendos proyectos de I+D+i (del MEC y del MICINN: SEJ2004-01098 y CSO2008-03337), dirigidos ambos por Benjamín González Rodríguez, coautor de este artículo.
Cada vez son más los artículos que consideran la opinión ciudadana como información esencial para... more Cada vez son más los artículos que consideran la opinión ciudadana como información esencial para evaluar la acción administrativa local. La investigación reciente viene demostrando la presencia de un vínculo empírico entre la calidad percibida de los servicios locales y la intención de conducta favorable al titular del gobierno local. Debido a que la literatura científica también acumula evidencia de relación entre ideología y preferencia política, el propósito de este artículo ha consistido en averiguar cómo interviene la ideología en el impacto de la calidad percibida sobre la preferencia política, es decir, si lo anula o lo modera y en qué grado. Como los conceptos calidad percibida e intención de conducta son constructos enlazados causalmente, el tratamiento ha requerido el empleo de la metodología de ecuaciones estructurales (SEM). El análisis progresa verificando las hipótesis de espuridad y moderación, pero los datos van a rechazarlas y apuntar a un mecanismo explicativo no esperado. Una encuesta realizada en nueve localidades de la provincia de Valencia (España) ha permitido comprobar el sorprendente papel que la ideología está desempeñando en la formación de la intención política del ciudadano a partir de las evaluaciones del servicio. Palabras clave: evaluación de los servicios; escala multidimensional jerárquica; calidad percibida; opinión pública; modelos de ecuaciones estructurales; análisis de invarianza; ideología política; Administración local; preferencia política.
Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005.
An experimental study of several decision issues for wrapper Feature Selection with Multi-Layer P... more An experimental study of several decision issues for wrapper Feature Selection with Multi-Layer Perceptrons is presented, namely the stopping criterion, the data set where the saliency is measured and the network retraining before computing the saliency. Experimental results with the Sequential Backward Selection procedure indicate that the increase in the computational cost associated with retraining the network with every feature temporarily removed before computing the saliency is rewarded with a significant performance improvement. Despite being quite intuitive, this idea has been hardly used in practice. Regarding the stopping criterion and the data set where the saliency is measured, the procedure profits from measuring the saliency in a validation set, as reasonably expected. A somehow non-intuitive conclusion can be drawn by looking at the stopping criterion, where it is suggested that forcing overtraining may be as useful as early stopping. 1 There is no commonly accepted definition of the relevance of a variable (see [3], [7], for example). Given a data set, we consider that a variable is irrelevant for a Supervised Machine Learning system when its optimal performance is not affected negatively by the absence of that variable ([7], page 29). Note that this is a dynamic definition, since the relevance of a variable may be affected by the presence or absence of other ones.
Uploads
Papers by Enrique Romero