G. Astray

Evaluation of atmospheric Poaceae pollen concentration using a neural network applied to a coastal Atlantic climate region

In the South of Europe an important percentage of population suffers pollen allergies, being the Poaceae pollen the major source. One of aerobiologys objectives is to develop statistical models enabling the short- and long-term prediction of atmospheric pollen concentrations to take preventative measures to protect allergic patients from the severity of the atmospheric pollen season. The implementation of a computational model based on supervised MLP neural network was applied for the prediction of the atmospheric Poaceae pollen concentration. There is a good correlation between the values predicted by the ANN for the training cases in comparison with the real pollen concentrations. A high coefficient of linear regression (R(2)) of 0.9696 was obtained. The accuracy of the neural network developed was tested with data from 2006 and 2007, which was not taken into account to establish the aforementioned models. Neural networks provided us a good tool to forecasting allergenic airborne pollen concentration helping the automation of the prediction system in the aerobiological information diffusion to the population suffering from allergic problems.

The use of artificial neural networks to forecast biological atmospheric allergens or pathogens only as Alternaria spores

The monitoring of atmospheric Alternaria spores is of major importance due to their adverse effects on crops and their role as human allergens. Most species act as plant pathogens, prompting considerable economic losses worldwide on important crops such as potato, tomato or wheat. Fungal spores can also have serious detrimental effects on human health, triggering respiratory diseases and allergenic processes. The aim of this study was not only to examine the relationship between the atmospheric Alternaria spore content and the prevailing meteorological parameters, but also to predict the atmospheric Alternaria spore content in the Northwest Spain using a novel data analysis technique, ANNs (Artificial Neural Networks). A Hirst-type LANZONI VPPS 2000 volumetric 7-day recording sampler was used to collect the airborne spores from 1997 to 2008. Neural networks provided us with a good tool for forecasting Alternaria airborne spore concentration, and thus could help the automation of the prediction system in the aerobiological information diffusion to the population suffering from allergic problems or the prevention of considerable economic worldwide losses on important crops. Our proposed model would be applied to different geographical areas; nevertheless, the adjustment of the model, by using the available and adequate variables, would be realised in each case.

Organic Reactivity in Aot-Stabilized Microemulsions

Microemulsions are highly versatile reaction media, which currently find many applications. In this review, we shall describe recent trends in the use of microemulsions as organic reaction media, and present models for their functioning, in particular the pseudophase model. This model allows a quantitative explanation of organic reactivity in these microheterogeneous media.

Cyclodextrin-Surfactant Mixed Systems as Reaction Media

In recent years our reseach group has investigated the chemical behaviour of β-cyclodextrin (CD)/surfactant mixed systems and their characteristics as reaction media. The results have been interpreted in terms of a pseudophase model that takes into account the formation of both CD-surfactant and CD-substrate complexes and also, in some cases, the exchange of X- and OH- ions between the micellar and aqueous pseudophases. from the experimental results it was concluded that the presence of CD has no effect on existing micelles but raises the critical micellar concentration (cmc). on the other hand, at surfactant concentrations above the cmc, competition between the micellisation and complexation processes leads to the existence of a significant concentration of free CD in equilibrium with the micellar aggregates. The percentage of uncomplexed β-CD in equilibrium with the micellar system increases on increasing the hydrophobicity of the surfactant molecule. This behaviour was justified taking into account the existence of two simultaneous processes: complexation of surfactant monomers by CD and the process of self-assembly to form micellar aggregates. The autoaggregation of surfactant monomers is more important than the complexation process in this mixed system. Varying the hydrophobicity of the surfactant monomer enabled us to determine that the percentages of uncomplexed CD in equilibrium with the micellar system were in the range of 5-95%. When the surfactant self-assembly structure is a vesicle, the free CD in the CD/surfactant mixed system yields a percentage of 100%.

Esters flash point prediction using artificial neural networks

In this article, an artificial neural network to predict the flash point of 95 esters was implemented. Four variables were used for its development. A neural network with 4‐5‐8‐5‐1 topology was encountered to gain the best agreement of the experimental results with those predicted (square correlation coefficient (R2) and root mean square error were 0.99 and 5.46 K for the training phase and 0.96 and 13.02 K for the testing set).

Artificial Intelligence for Electrical Percolation of AOT-based Microemulsions Prediction

Abstract Different Artificial Neural Network architectures have been assayed to predict percolation temperature of AOT/i-C8/H2O microemulsions. A Perceptron Multilayer Artificial Neural Network with five entrance variables (W value of the microemulsions, additive concentration, molecular weight of the additive, atomic radii and ionic radii of the salt components) was used. Best ANN architecture was formed by five input neurons, two middle layers (with eleven and seven neurons respectively) and one output neuron. Root Mean Square Errors (RMSEs) are 0.18°C (R = 0.9994) for the training set and 0.64°C (R = 0.9789) for the prediction set.

Influence Prediction of Small Organic Molecules (Ureas and Thioureas) Upon Electrical Percolation of AOT-Based Microemulsions Using Artificial Neural Networks

Abstract In order to predict percolation temperature of AOT-Based microemulsions (AOT/iC8/H2O w/o microemulsions) in the presence of small organic molecules (ureas and thioureas), different Artificial Neural Network architectures (ANN) have been carried out using a Perceptron Multilayer Artificial Neural Network with three entrance variables (W = value of the microemulsion, additive concentration, logP value). Best ANN architecture consists in three input neurons, one middle layer (with two neurons) and one output neuron. Correlation values were R = 0.9251 for the training set and R = 0.9719 for the prediction set.

Artificial neural networks: a promising tool to evaluate the authenticity of wine Redes neuronales: una herramienta prometedora para evaluar la autenticidad del vino

Artificial Neural Networks (ANNs) have demonstrated to be a good tool to characterise, model and predict a great quantity of non-linear processes. In this article, we have used ANNs in the classification of different wine-making processes of the variety Vinhão (Vitis vinifera) for crops between the years 2000 and 2004. After being trained employing the data corresponding to years from 2000 to 2004, the ANNs demonstrated a root mean square error (RMSE) index between the real data and the calculated ones always lower than 0.14. Furthermore, their operation has been verified by using the previously reserved data of 10 famous wines. As a result, a RMSE index between observed and calculated data always lower than 0.17 was obtained for all of them, confirming the capacity of the ANN as a model of prediction of wine processes for this variety.

Alkaline fading of triarylmethyl carbocations in self-assembly microheterogeneous media

This review reports on the alkaline fading of crystal violet and other related carbocations in the presence of different microheterogeneous media (micelles, microemulsions and vesicles).

Percolation Threshold of AOT Microemulsions with n-Alkyl Acids as Additives Prediction by Means of Artificial Neural Networks

Abstract Different artificial neural networks architectures have been assayed to predict percolation temperature of AOT/iC8/H2O microemulsions in the presence of n-alkyl acids with a chain length between 0 and 24 carbons, using a multilayer perceptron with five easy-acquired entrance variables (number of carbons, log P, length of the hydrocarbon chain, pKa and acid concentration). The evaluation of the neural networks was carried out by means of RMSE and IDP, resulting that the architecture with better results consists in five input neurons, two middle layers (with five and ten neuron respectively) and one output neuron. Results prove that Artificial Neural Networks are a useful tool elaborating models to predict percolation temperature of microemulsion systems in the presence of additives.