Cortisol, DHEA, performance and training in elite swimmers

Salivary cortisol (C) and DHEA concentrations were measured in 9 elite swimmers (4 female and 5 male) over a 37-week period, 5 to 12 times per swimmer, before 68 competitions. For female and male swimmers, no significant relationship was found between C, DHEA and performance. For the whole group, C was negatively correlated with week number of training (r = -0.31, p < 0.01). The incorporation of the cumulated distance swum as a second variable in the regression increased r to 0.56 (p < 0.01). The higher the cumulated distance swum, the higher C. No significant relationship was found between DHEA and distance swum. For individual swimmers, 3 of 4 females showed a significant negative relationship between C and cumulated dry-land training. No equivalent relationship was found for DHEA. The 2 males practicing dry-land training showed a significant and negative relationship between DHEA and cumulated dry-land training. No equivalent relationship was found for C. Thus, C and DHEA were not good predictors of swimming performance. C for individual females, and DHEA for individual males were considered useful markers for dry-land training stress.

I-Spi: innovative shared practical ideas: a support guide for literacy and numeracy in the early years of schooling, for home tutor/supervisors in distance education

Innovative Shared Practical Ideas (I-Spi) is a guide to help you and your children learn together. It is designed to affirm, support and strengthen your role as home tutor/supervisors in your daily learning sessions with your children. In this guide particular emphasis is given to the value of talk, formal and informal early literacy and numeracy practices (including ideas from distance school lessons, from home tutor/supervisors, research, and beyond), assessment of these practices together with informal assessment ideas for gauging your children’s literacy and numeracy progress, and stepping in and building on strategies

Improving support vector solutions by selecting a sequence of training subsets

In this paper we demonstrate that it is possible to gradually improve the performance of support vector machine (SVM) classifiers by using a genetic algorithm to select a sequence of training subsets from the available data. Performance improvement is possible because the SVM solution generally lies some distance away from the Bayes optimal in the space of learning parameters. We illustrate performance improvements on a number of benchmark data sets.