Influence of Training Set Selection in Artificial Neural Network-Based Propagation Path Loss Predictions

This paper analyzes the use of artificial neural networks (ANNs) for predicting the received power/path loss in both outdoor and indoor links. The approach followed has been a combined use of ANNs and ray-tracing, the latter allowing the identification and parameterization of the so-called dominant path. A complete description of the process for creating and training an ANN-based model is presented with special emphasis on the training process. More specifically, we will be discussing various techniques to arrive at valid predictions focusing on an optimum selection of the training set. A quantitative analysis based on results from two narrowband measurement campaigns, one outdoors and the other indoors, is also presented.

Assessment of phytostabilization of Pb/Zn mine soils with organic amendments and a native metallicolous ecotype of Festuca Rubra l.: implications for amendment selection and recovery of soil health

La contaminación del suelo es una de las principales amenazas para los ecosistemas y la salud humana. Actualmente, desde un punto de vista tanto económico como ambiental, la fitoestabilización es la mejor tecnología para remediar suelos contaminados con elevadas concentraciones de metales como son los suelos mineros. La fitoestabilización asistida consiste en el empleo de plantas y enmiendas orgánicas y/o inorgánicas con el fin de reducir la movilidad y la biodisponibilidad de los contaminantes y recuperar la salud de suelo. En este trabajo se han realizado ensayos en microcosmos y en campo centrándonos en la salud del suelo minero contaminado con Pb y Zn durante un proceso de fitoestabilización empleando enmiendas orgánicas (purines vacunos, gallinaza, estiércol de oveja y lodos de papelera mezclados con gallinaza) y/o la especie metalífera Festuca rubra con el objetivo de (i) estudiar las interacciones suelo-enmienda responsables de los cambios inducidos por el proceso de quimioestabilización en las propiedades físicoquímicas y biológicas del suelo, (ii) evaluar la efectividad del proceso de fitoestabilización sobre suelos vegetados y de la revegetación sobre suelos desnudos (iii) valorar la idoneidad de distintos indicadores químicos y biológicos (parámetros microbianos y de la vegetación) para monitorizar la efectividad de la fitoestabilización asistida en términos de reducción de la biodisponibilidad de metales en el suelo, mejora de la vegetación y de la recuperación de la salud del suelo. La aplicación de enmiendas al suelo minero supone una entrada de materia orgánica y nutrientes que conduce a una disminución de la biodisponibilidad de metales, facilitando la colonización de las plantas y el crecimiento de la vegetación nativa, además de estimular la actividad microbiana del suelo. El pH del suelo es un factor crítico que condiciona la movilidad de los metales y la toxicidad del suelo. Las poblaciones microbianas de las enmiendas no modificaron la diversidad funcional de las comunidades microbianas nativas de la mina. Los purines vacunos y los lodos de papelera mezclados con gallinaza son los tratamientos más efectivos en el proceso de fitoestabilización asistida bajo condiciones de campo. La gallinaza fue el tratamiento que más estimuló el crecimiento de la vegetación nativa y la colonización en los suelos desnudos. El bioensayo de elongación radical de lechuga es un test sensible, sencillo y barato para evaluar la biodisponibilidad de metal y la ecotoxicidad del suelo. Los tocoferoles son biomarcadores de exposición a metales con potencial para su implementación en bioensayos de toxicidad. Este trabajo permite concluir que la población metalífera de F. rubra, combinada con enmiendas orgánicas, es una excelente candidata para los proyectos de fitoestabilización asistida. Además, la monitorización simultánea de los parámetros fisicoquímicos y microbiológicos del suelo y de su ecotoxicidad permite una evaluación adecuada de la salud del suelo, así como la selección de enmiendas apropiadas para el desarrollo de un proceso fitoestabilizador.

On the Selection of Non-Invasive Methods Based on Speech Analysis Oriented to Automatic Alzheimer Disease Diagnosis

The work presented here is part of a larger study to identify novel technologies and biomarkers for early Alzheimer disease (AD) detection and it focuses on evaluating the suitability of a new approach for early AD diagnosis by non-invasive methods. The purpose is to examine in a pilot study the potential of applying intelligent algorithms to speech features obtained from suspected patients in order to contribute to the improvement of diagnosis of AD and its degree of severity. In this sense, Artificial Neural Networks (ANN) have been used for the automatic classification of the two classes (AD and control subjects). Two human issues have been analyzed for feature selection: Spontaneous Speech and Emotional Response. Not only linear features but also non-linear ones, such as Fractal Dimension, have been explored. The approach is non invasive, low cost and without any side effects. Obtained experimental results were very satisfactory and promising for early diagnosis and classification of AD patients.

Countervailing incentives in adverse selection models. A synthesis

In this paper we propose a simple method of characterizing countervailing incentives in adverse selection problems. The key element in our characterization consists of analyzing properties of the full information problem. This allows solving the principal problem without using optimal control theory. Our methodology can also be applied to different economic settings: health economics, monopoly regulation, labour contracts, limited liabilities and environmental regulation.

Autophagic Marker MAP1LC3B Expression Levels Are Associated with Carotid Atherosclerosis Symptomatology

Objectives: The mechanism by which atheroma plaque becomes unstable is not completely understood to date but analysis of differentially expressed genes in stable versus unstable plaques may provide clues. This will be crucial toward disclosing the mechanistic basis of plaque instability, and may help to identify prognostic biomarkers for ischaemic events. The objective of our study was to identify differences in expression levels of 59 selected genes between symptomatic patients (unstable plaques) and asymptomatic patients (stable plaques). Methods: 80 carotid plaques obtained by carotid endarterectomy and classified as symptomatic (>70% stenosis) or asymptomatic (>80% stenosis) were used in this study. The expression levels of 59 genes were quantified by qPCR on RNA extracted from the carotid plaques obtained by endarterectomy and analyzed by means of various bioinformatic tools. Results: Several genes associated with autophagy pathways displayed differential expression levels between asymptomatic and symptomatic (i.e. MAP1LC3B, RAB24, EVA1A). In particular, mRNA levels of MAP1LC3B, an autophagic marker, showed a 5-fold decrease in symptomatic samples, which was confirmed in protein blots. Immune system-related factors and endoplasmic reticulum-associated markers (i.e. ERP27, ITPR1, ERO1LB, TIMP1, IL12B) emerged as differently expressed genes between asymptomatic and symptomatic patients. Conclusions: Carotid atherosclerotic plaques in which MAP1LC3B is underexpressed would not be able to benefit from MAP1LC3B-associated autophagy. This may lead to accumulation of dead cells at lesion site with subsequent plaque destabilization leading to cerebrovascular events. Identified biomarkers and network interactions may represent novel targets for development of treatments against plaque destabilization and thus for the prevention of cerebrovascular events.

Feature Selection for Speech Emotion Recognition in Spanish and Basque: On the Use of Machine Learning to Improve Human-Computer Interaction

Study of emotions in human-computer interaction is a growing research area. This paper shows an attempt to select the most significant features for emotion recognition in spoken Basque and Spanish Languages using different methods for feature selection. RekEmozio database was used as the experimental data set. Several Machine Learning paradigms were used for the emotion classification task. Experiments were executed in three phases, using different sets of features as classification variables in each phase. Moreover, feature subset selection was applied at each phase in order to seek for the most relevant feature subset. The three phases approach was selected to check the validity of the proposed approach. Achieved results show that an instance-based learning algorithm using feature subset selection techniques based on evolutionary algorithms is the best Machine Learning paradigm in automatic emotion recognition, with all different feature sets, obtaining a mean of 80,05% emotion recognition rate in Basque and a 74,82% in Spanish. In order to check the goodness of the proposed process, a greedy searching approach (FSS-Forward) has been applied and a comparison between them is provided. Based on achieved results, a set of most relevant non-speaker dependent features is proposed for both languages and new perspectives are suggested.

Contributions to learning Bayesian network models from weakly supervised data: Application to Assisted Reproductive Technologies and Software Defect Classification

162 p.

The Interplay between Natural Selection and Susceptibility to Melanoma on Allele 374F of SLC45A2 Gene in a South European Population

We aimed to study the selective pressures interacting on SLC45A2 to investigate the interplay between selection and susceptibility to disease. Thus, we enrolled 500 volunteers from a geographically limited population (Basques from the North of Spain) and by resequencing the whole coding region and intron 5 of the 34 most and the 34 least pigmented individuals according to the reflectance distribution, we observed that the polymorphism Leu374Phe (L374F, rs16891982) was statistically associated with skin color variability within this sample. In particular, allele 374F was significantly more frequent among the individuals with lighter skin. Further genotyping an independent set of 558 individuals of a geographically wider population with known ancestry in the Spanish population also revealed that the frequency of L374F was significantly correlated with the incident UV radiation intensity. Selection tests suggest that allele 374F is being positively selected in South Europeans, thus indicating that depigmentation is an adaptive process. Interestingly, by genotyping 119 melanoma samples, we show that this variant is also associated with an increased susceptibility to melanoma in our populations. The ultimate driving force for this adaptation is unknown, but it is compatible with the vitamin D hypothesis. This shows that molecular evolution analysis can be used as a useful technology to predict phenotypic and biomedical consequences in humans.