Hierarchical Learning in Polynomial Support Vector Machines

We study the typical properties of polynomial Support Vector Machines within a Statistical Mechanics approach that takes into account the number of high order features relative to the input space dimension. We analyze the effect of different features' normalizations on the generalization error, for different kinds of learning tasks. If the normalization is adequately selected, hierarchical learning of features of increasing order takes place as a function of the training set size. Otherwise, the performance worsens, and there is no hierarchical learning at all.     

Effect of impregnation on the degree of polymerization of cellulose during bleaching of sodapulped sugar cane bagasse

Samples of sugar cane bagasse pulps, digested at different times without and with previous impregnation, have been bleached and the bleached materials submitted to a detailed study of their degree of polymerization (DP) and the corresponding distributions. The respective results turn out that like in the case of the effect of impregnation on the digestion process, impregnation affects also the bleaching reaction of the digested samples, causing a further remarkable decrease of the degree of polymerization of the cellulose molecules and a deviation of its distribution from the character of a most probable one. The impregnated bleached materials show in addition to the decrease of their DP a certain, albeit slight, diminution in their tensile strength.     

Estimation of potentially available protein in infant starting formulas for term and preterm neonates

Sixteen milk-based starting formulas were analyzed with the aim of calculating their "true protein" content and assessing "in vitro" protein digestibility, in order to estimate levels of potentially available protein. Ten of them were designed for term infants: 7 had a casein:whey protein ratio 40:60 (adapted formulas) and 3 a ratio 80:20 (non-adapted); the 6 remaining formulas (all adapted) were for preterm infants. Nitrogen was determined by the Kjeldahl method. True protein was calculated as (total N--non-protein N) x 6.25. NPN was determined in the soluble fraction, after protein precipitation with 24% trichloroacetic acid and centrifugation. Digestibility was assessed by digestion with pepsin and pancreatin, and defined as the increase in non-protein N after enzymatic digestion. Values for true protein were from 1.3 to 2.3 g/dL and for non protein N, from 4.5 to 13.7%. Digestibility values varied between 59.0 and 92.5%; an inverse trend was observed between protein digestibility and protein content. Considering both the "true protein" levels and their digested proportions, all preterm and 60% of the term formulas would present potentially available amounts below those recommended. These observations constitute an alert, even though this method of assessing "in vitro" protein digestibility represents only an approximation to physiological processes; however, it could be useful in order to evaluate the intensity of the heat treatments to which these formulas were subjected. On the other hand, since NPN allows the estimation of the true protein provided by the formulas, either its percentage or the true protein content could be included on the label.     

Characterization of the Sonar Signals Benchmark

We study the classification of sonar targets first introduced by Gorman & Sejnowski (1988). We discovered that not only the training set and the test set of this benchmark are both linearly separable, although by different hyperplanes, but that the complete set of patterns, training and test patterns together, is also linearly separable. The distances of the patterns to the separating hyperplane determined by learning with the training set alone, and to the one determined by learning the complete data set, are presented.     

Waterborne polyurethane nanocomposites based on vegetable oil and microfibrillated cellulose

This work analyzes the differences in the final properties of two waterborne polyurethanes (WBPU) prepared with two macrodiols of different chemical structure, but similar molecular weight, as well as the variations caused by incorporating low percentages of microfibrillated cellulose nanocrystals. One of the polyurethanes was based on a synthetic but biodegradable precursor (polycaprolactone diol, PCL) and a second one based on a bio‐based macrodiol derived from castor oil (CO1). The bio‐based material presented higher mechanical properties at room temperature than the synthetic one, with the Young's modulus (MPa) ranging from 2.23 ± 0.09 to 84.88 ± 0.96 for the PCL and bio‐based WBPUs, respectively. Additionally, the PCL‐based WBPU showed to be more sensitive to the incorporation of cellulose than the bio‐based WBPU, and it also suffered changes during time due to delayed crystallization. The behavior of the two systems were compared and related to the different structure of the macrodiols that led to different interfacial interactions. © 2016 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2016 , 133, 44207.     

On the existence of the nonlinear weighted least squares estimate for a three-parameter Weibull distribution

The problem of nonlinear weighted least squares fitting of the three-parameter Weibull distribution to the given data (w_i,t_i,y_i), i=1,…,n, is considered. The part w_i>0 of the data stands for the data weights. It is shown that the best least squares estimate exists provided that the data satisfy just the following two natural conditions: (i) 0<t₁<t₂<?<t_n and (ii) 0<y₁<y₂<?<y_n<1. To support this, an illustrative numerical example is given.     

Comparison procedure of predicting the time to default in behavioural scoring

The paper deals with the problem of predicting the time to default in credit behavioural scoring. This area opens a possibility of including a dynamic component in behavioural scoring modelling which enables making decisions related to limit, collection and recovery strategies, retention and attrition, as well as providing an insight into the profitability, pricing or term structure of the loan. In this paper, we compare survival analysis and neural networks in terms of modelling and results. The neural network architecture is designed such that its output is comparable to the survival analysis output. Six neural network models were created, one for each period of default. A radial basis neural network algorithm was used to test all six models. The survival model used a Cox modelling procedure. Further, different performance measures of all models were discussed since even in highly accurate scoring models, misclassification patterns appear. A systematic comparison ‘3 + 2 + 2’ procedure is suggested to find the most effective model for a bank. Additionally, the survival analysis model is compared to neural network models according to the relative importance of different variables in predicting the time to default. Although different models can have very similar performance measures they may consist of different variables. The dataset used for the research was collected from a Croatian bank and credit customers were observed during a 12-month period. The paper emphasizes the importance of conducting a detailed comparison procedure while selecting the best model that satisfies the users’ interest.     

Poly(methyl methacrylate)‐modified vinyl ester thermosets: Morphology,volume shrinkage,and mechanical properties

Thermoset materials obtained from styrene/vinyl ester resins of different molecular weights modified with poly(methyl methacrylate) (PMMA) were prepared and studied. Scanning electron microscopy and transmission electron microscopy micrographs of the fracture surfaces allowed the determination of a two‐phase morphology of the modified networks. Depending on the molecular weight of the vinyl ester oligomer, the initial content of the PMMA additive, and the selected curing temperature, different morphologies were obtained, including the dispersion of thermoplastic‐rich particles in a thermoset‐rich matrix, cocontinuous structures, and the dispersion of thermoset‐rich particles in a thermoplastic‐rich matrix (phase‐inverted structure). Density measurements were performed to determine the effect of the PMMA‐modifier concentration and curing temperature on the volume shrinkage of the final materials. The development of cocontinuous or thermoplastic‐rich matrices was not too effective in controlling the volume shrinkage of the studied vinyl ester systems. The evaluation of the dynamic mechanical behavior, flexural modulus, compressive yield stress, and fracture toughness showed that the addition of PMMA increased the fracture resistance without significantly compromising the thermal or mechanical properties of the vinyl ester networks. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci 2007     

Changes in food availability in metropolitan Santiago Chile according to income (quintiles) 1988-1997

Changes in household food availability from 1988-1997 obtained in a representative sample of Metropolitan Santiago were assessed; and related to observed changes in the epidemiological profile. We evaluated expenditures in food and beverages from the IV and V Household Expenditure Surveys conducted every 10 years by the National Institute of Statistics to calculate the Consumer Price Index. Food items were similarly grouped and expenditures from both surveys adjusted to concordance by assigning prices according to of food; the units consumed outside and at home were determined. Food expenditures increased, leading to greater food availability in all households and income categories; the increment was largest in the poorest quintiles. Apparent consumption of processed cereals, pastries and baked goods, poultry and pork, processed dairies, beverages and juices, dressings and mayonnaise, pre cooked meals and meals consumed "out of home" increased. Fish, vegetables and fruits increased slightly with a concomitant decrease in legumes. The present dietary pattern is also characterized by a greater availability of animal food products and processed foods; increased energy density, fat and saturated fat energy, sugars and high glycemic index foods; lower in phytochemicals, antioxidants, dietary fiber and omega-3 fatty acids. These changes are probably having a significant impact on the epidemiological and nutritional profile of the country; as manifested by an epidemic increase in obesity and chronic disease burden related to diet and physical inactivity.