An Iterative Method for Deciding SVM and Single Layer Neural Network Structures

We present two new classifiers for two-class classification problems using a new Beta-SVM kernel transformation and an iterative algorithm to concurrently select the support vectors for a support vector machine (SVM) and the hidden units for a single hidden layer neural network to achieve a better generalization performance. To construct the classifiers, the contributing data points are chosen on the basis of a thresholding scheme of the outputs of a single perceptron trained using all training data samples. The chosen support vectors are used to construct a new SVM classifier that we call Beta-SVN. The number of chosen support vectors is used to determine the structure of the hidden layer in a single hidden layer neural network that we call Beta-NN. The Beta-SVN and Beta-NN structures produced by our method outperformed other commonly used classifiers when tested on a 2-dimensional non-linearly separable data set.     

Iron incorporated rice husk silica as a sorbent for hexavalent chromium attenuation in aqueous system

An environmentally benign metal oxide, Fe, was incorporated into silica derived from rice husk via the sol gel route and the physiognomies of both the modified (IRS) and raw rice husk derived silica (RHS) were studied via FTIR and XRD analysis and pH_PZC and surface area determinations. The stability of the Fe incorporated into the silica matrix, determined via the toxicity characteristic leaching procedure, showed that the integrity of the sorbent was intact only in basic medium but got vitiated in acidic medium. The sorption process conformed to the pseudo second order model than reversible first order and pseudo first order kinetic models and the rate of sorption of Cr(VI) onto either sorbents was determined by film diffusion. Process variables optimization showed that the amount of Cr(VI) removed per gram of sorbent reduced with increase in initial solution pH and the negative impact of the anionic interference was more in the presence of SO₃²⁻ than NO₃²⁻ and Cl⁻. The predicted Langmuir monolayer sorption capacity (mg/g) of the IRS (63.69) was higher than that of the RHS (61.35). The value of the mean free energy (kJ/mol) of sorption, obtained for IRS (267.26) and RHS (100.00), and the significant changes in the peak positions of specific functional groups on the Cr(VI) laden sorbents showed that chemisorption was the dominant mechanism of Cr(VI) uptake.     

A bipolar implementation of the Beta neuron

The great flexibility of the Beta function and its universal approximation characteristics, make Beta basis function neural networks (BBFNNs) very useful. We present a hardware implementation of the Beta neuron. The proposed circuit was designed by using a standard bipolar technology. PSPICE simulations show the good concordance of the output of our circuit with the analytic Beta function. We also successfully integrated the electronic Beta neuron in the design of a BBFNN that approximates a nonlinear mapping.     

A Beta basis function neural network in CMOS subthreshold mode

Beta basis function neural networks (BBFNNs) are powerful systems for learning and universal approximation. In this paper, we present a hardware implementation of the Beta neuron using the CMOS subthreshold mode. We describe the low power–low voltage analogue Beta neuron circuit. Three main modules are used to realize the electronic Beta function: a logarithmic currentto-voltage converter, a multiplier and an exponential voltage-to-current converter. Simulation results show the validity of our neural hardware implementation. The parameters of the electronic Beta function are controlled independently by current sources. This analogue implementation could be used easily to realize analogue BBFNNs.     

Surgery of the perforating veins of the leg: open and endoscopic techniques]

Surgical treatment of incompetent perforating veins of the leg is best carried out through a sub-aponeurotic approach after localisation by coloured Duplex-Scanning. Open techniques are long established. When there is only one perforator, a short 5 cm approach is centred over the previously localised perforator. In the presence of multiple perforators, an extensive approach is necessary. This usually means a single long incision, situated para-tibially in the Linton approach, postero-medially in the Dodd approach or posteriorly in the Felder approach. The advantage of the last two approaches is that they are situated at a distance from any eventual cutaneous lesion. The approach can also be through several separate oblique incisions (De Palma). More recently introduced, endoscopic techniques have the advantage of reducing the incidence of cutaneous complications such as necrosis or infections and of shortening the duration of hospitalisation. Some authors work with re-usable instruments. The perforators are divided after coagulation (Hauer, Sattler). Others use disposable instruments, insufflate the sub-aponeurotic space and clip the perforators to prevent bleeding (Gloviczki). These techniques are currently undergoing evaluation.     

Utilisation des sondes TDR pour estimer la masse de soluté retenu par une couche d'argile lors d'une infiltration

Résumé. Un essai de longue durée est réalisé sur un matériau d'étanchéité, l'argile du Gault. Par une infiltration continue du lixiviat de scories, sous une charge hydraulique constante de 98,5 cm, nous avons suivi la variation spatio-temporelle de l'humidité θ (z, t) et de la conductivité électrique σ (z, t) dans une colonne du sol. La relation entre la teneur en eau volumique (θ) et la constante diélectrique (ε) du sol est établie en utilisant la méthode TDR. Nous avons également établi une relation entre la fraction soluble de la solution et sa conductivité électrique. Cette relation permet de suivre la distribution spatio-temporelle du soluté dans la couche de sol. Plusieurs chercheurs ont montré qu'il est possible de tracer les courbes d'élution d'un traceur non réactif en utilisant la méthode TDR. Cette méthode permet alors d'éviter l'utilisation des bougies poreuses d'extraction des solutions du sol et les analyses chimiques co?teuses de l'éluant. Les résultats des essais d'infiltration effectués sur la couche d'argile montrent que la conductivité électrique mesurée à l'aide des sondes TDR sert à localiser la fraction soluble dans la colonne. Le coefficient de dispersion calculé est resté constant au cours du temps autour d'une valeur moyenne de 1,5×10^–10 m²/s. Ce qui montre que le coefficient de dispersion n'est pas très influencé par le sol et par son degré de saturation. Electronic Publication