Modification of wool fiber using steam explosion

Wool fiber was modified by steam explosion in this study. SEM results show that some scales on the fiber surface were cleaved and tiny grooves generated during the explosion. FTIR results suggest no evident changes in the chemical composition of the fiber after the explosion treatment. However, the crystallinity of the fiber decreased slightly as the steam pressure increased based on the X-ray results. In the thermal analysis, DSC results show that the temperature corresponding to vaporization of absorbed water and cleavage of disulfide bonds respectively decreased as the steam pressure increased. The reduction in thermal decomposition energy of the treated fiber indicates that steam explosion might have destroyed some crystals and crosslinks of macromolecular chains in the fiber. The treatment also led to some alterations of the fiber properties, including reduction in strength, moisture regain and solubility in caustic solution.

A meta-learning approach to automatic kernel selection for support vector machines

Appropriate choice of a kernel is the most important ingredient of the kernel-based learning methods such as support vector machine (SVM). Automatic kernel selection is a key issue given the number of kernels available, and the current trial-and-error nature of selecting the best kernel for a given problem. This paper introduces a new method for automatic kernel selection, with empirical results based on classification. The empirical study has been conducted among five kernels with 112 different classification problems, using the popular kernel based statistical learning algorithm SVM. We evaluate the kernels’ performance in terms of accuracy measures. We then focus on answering the question: which kernel is best suited to which type of classification problem? Our meta-learning methodology involves measuring the problem characteristics using classical, distance and distribution-based statistical information. We then combine these measures with the empirical results to present a rule-based method to select the most appropriate kernel for a classification problem. The rules are generated by the decision tree algorithm C5.0 and are evaluated with 10 fold cross validation. All generated rules offer high accuracy ratings.

Lupin (Lupinus albus) protein isolate (L-ISO) has adequate nutritional value and reduces large intestinal weight in rats after restricted and ad libitum feeding.

Background: A protein isolate from white lupin (Lupinus albus; L-ISO) has potential as a novel human food ingredient, but its nutritional effects are unknown.

Methods: We evaluated protein quality and effects on body composition in rats of isoenergic diets of L-ISO, lactalbumin, or casein with both restricted (10-day) and ad libitum (28-day)intake. The diets were equivalent in protein per se, but supplementation was used to balance essential amino acid levels.

Results: In both studies, the rats consumed similar amounts of each diet, and no effect of diet on the gain:feed ratio was observed--though gain:N ratio and net protein utilization were slightly lower for the L-ISO diet. Lower large intestinal weights after the L-ISO than after the lactalbumin diet were observed in both studies. The L-ISO diet resulted in lowered body fat percentage in the 10-day study but in an elevated level in the 28-day study. Liver composition (DNA, RNA, glycogen, and fat) and plasma levels of some amino acids (His, Thr, Ala, Pro, Tyr, Val and Met) were affected by diet, but no effects on plasma lipid, glucose, or uric acid were observed.

Conclusion: The L-ISO diet did not affect feed intake and has adequate nutritional quality in rats whilst modifying large intestinal weight in a potentially beneficial manner--suggesting potential for this protein in human nutrition.