Word sense disambiguation for spam filtering

Spam has become a major issue in computer security because it is a channel for threats such as computer viruses, worms, and phishing. More than 86% of received e-mails are spam. Historical approaches to combating these messages, including simple techniques such as sender blacklisting or the use of e-mail signatures, are no longer completely reliable. Many current solutions feature machine-learning algorithms trained using statistical representations of the terms that most commonly appear in such e-mails. However, these methods are merely syntactic and are unable to account for the underlying semantics of terms within messages. In this paper, we explore the use of semantics in spam filtering by introducing a pre-processing step of Word Sense Disambiguation (WSD). Based upon this disambiguated representation, we apply several well-known machine-learning models and show that the proposed method can detect the internal semantics of spam messages.     

Carbon membranes from cellulose and metal loaded cellulose

The focus of this work was to find a low-cost precursor for carbon molecular sieve (CMS) membranes, and a simple way of producing them. In addition, several ways of modifying a carbon material are described. The modification method used in this study was metal doping of carbon. CMS membranes were formed by vacuum carbonization of cellulose and metal loaded cellulose. Metal additives include oxides of Ca, Mg, Fe(III) and Si, and nitrates of Ag, Cu and Fe(III).The carbon membrane containing Fe-nitrate has promising separation performance for the gas pairs O₂/N₂ and CO₂/CH₄. Carbon containing nitrates of Cu or Ag show high selectivity, but reduced O₂ and CO₂ permeability compared to carbon with Fe-nitrate. Element analysis indicates that Cu migrates to the carbon surface, creating an extra layer resistance to gas transport. A silver mirror is also seen on the surface of Ag-nitrate-containing carbon. However, the Ag- and Cu-containing membranes show a high H₂ permeability. Adding metal oxides makes the carbon membranes retard the transport of easily condensable gases (e.g. CO₂). This can be exploited for enhanced H₂/CO₂ separation efficiency.     

The adsorption of mercuric ions by chitosan

The adsorption of mercuric ions by chitosan was investigated. The study of the adsorption kinetics shows that the rate of adsorption of mercuric ions on chitosan can be interpreted in terms of intraparticle diffusion as the rate-limiting step. The experimental data of adsorption equilibrium from mercuric chloride solutions correlate well with the Langmuir isotherm equation, although at high-solute concentrations, a multilayer type of adsorption with the subsequent increase in the uptake is observed. Column experiment confirms the ability of chitosan for the removal of mercuric ions from solutions in the absence of a high concentration of chlorides.     

Functional properties of a protein product from Caryodendron orinocense (Barinas nut)

The functional properties of Caryodendron orinocense protein product were investigated and compared with those of soybean (Glycina maxima). The product protein content was 24.47 g/100 g (Nx6.25). Solubility increased at both sides of the isoelectric point (pH 4.0) and with increased NaCl concentration up to 0.5M. Compared with soybean flour (50% protein), the protein product exhibited higher water and oil absorption, but lower emulsifying activity, emulsion stability, foaming capacity, and foam stability, the last one increase at higher pH. Emulsifying activity, foaming capacity, and foam stability were ionic strength dependent. C. orinocense protein product increased its emulsifying activity steadily from 0.05M to 0.75M NaCl, while it remained almost constant for soybean flour. Foaming capacity increased drastically at pH 10. The minimum time and concentration to form a gel was 20% in 4 min and 10% in 8 min for the Caryodendron protein product and soybean flour, respectively. The bulk density was 0.5056+/-0.0041 g/mL.     

Aspects of poly(p-phenylene sulfide) degradation and stabilization. Part 1. Influence of polymer end groups on exposure induced coloration

In this paper we reviewed some of our work on the photo induced coloration of Poly(p-phenylene sulfide) (PPS) and compared it to recent results obtained from the thermal treatment of resins that varied in molecular weight and ionic content. The results from this study suggest that coloration of amorphous PPS films after UV irradiation is caused by products different from those that cause coloration in semicrystalline PPS after thermoxidation. In addition, we found that polymer end group chemistry and resin molecular weights play important roles in the thermoxidative coloration of PPS, but only nominally influence the photooxidative coloration of the resin. PPS coloration may arise from the accumulation of specific, colored products (e.g., phenyl-phenyl moieties) and the possible presence of intermolecular aggregates. The mechanism and extent by which these structures occur in UV irradiated or heated PPS are likely dependent upon the polymer morphology of the samples tested.     

Dynamic behaviour of viologen-activated nanostructured TiO2: correlation between kinetics of charging and coloration

The dynamic response of viologen-activated nanostructured titanium dioxide has been studied by means of electrical and electro-optical measurements. We show that the state of charge of the semiconductor network is the key factor mediating between the electrode potential and colouration of viologen. Theoretically, we relate the electrode potential to the statistics of occupancy of both TiO₂ nanoparticles and oxidized viologen molecules attached to the surface, on the assumption of quasi-equilibrium of Fermi levels in these contacting phases. Experimentally, we determine the statistical function from steady-state measurements (electrochemical impedance spectroscopy) of the capacitance of the semiconductor film. From this understanding we explain the main features that correlate the simultaneous voltammetry and transmittance responses. Finally, the redox process of viologen is resolved separately from the TiO₂ response by means of transmittance data.     

Al ionic radius influence on molecular dynamics simulations ofγ-Al2O3

Molecular dynamics (MD) simulations of-Al₂O₃, using a pairwise additive interaction potential of Pauling's type with four different radii for aluminum atoms, were analysed in order to determine the influence of the radius on the modification of coordination numbers of A1 relative to the ideal structure, at two different temperatures of 300 and 1500 K. It is found that the best choice is the radius of penta-coordinated aluminum, reproducing structural and vibrational properties of the compound in excellent agreement with experimentally observed properties.     

Barocalorimetric technique for monitoring pressurized copolymer reactors

The problem of on‐line estimation of the conversion and composition evolutions in a pressurized batch copolymer reactor with temperature and pressure measurements was addressed. The estimation model consisted of mass and energy balances with a pressure equation built from phase‐equilibrium considerations. The application of a nonlinear geometric estimation approach yielded the underlying solvability condition with physical meaning, a straightforward estimator construction, and a conventional‐like tuning procedure. The resulting barocalorimetric estimator was an on‐line dynamic measurement processor with a model‐based predictor and a measurement‐driven corrector, and whose implementation did not require the polymerization rates and heat‐transfer coefficient function dependencies. The technique was tested with a representative laboratory styrene–butadiene system. © 2005 Wiley Periodicals, Inc. J Appl Polym Sci 96: 475–482, 2005     

Camera calibration in sport event scenarios

The main goal of this paper is the design of a novel and robust methodology for calibrating cameras from a single image in sport scenarios, such as a soccer field, or a basketball or tennis court. In these sport scenarios, the only references we use to calibrate the camera are the lines and circles delimiting the different regions. The first problem we address is the extraction of image primitives, including the challenging problems of shaded regions and lens distortion. From these primitives, we automatically recognise the location of the sport court in the scene by estimating the homography which matches the actual court with its projection onto the image. This is achieved even when only a few primitives are available. Finally, from this homography, we recover the camera calibration parameters. In particular, we estimate the focal length as well as the position and orientation in the 3D space. We present some experiments on models and real courts which illustrate the accuracy of the proposed methodology.