Conservation of water resources problems of small rivers in Russia

Translated from Gidrotekhnicheskoe Stroitel'stvo, No. 8, pp. 1–5, August, 1994.     

Determining hearing threshold from brain stem evoked potentials.Optimizing a neural network to improve classification performance

Feed-forward neural networks in conjunction with back-propagation are an effective tool to automate the classification of biomedical signals. Most of the neural network research to date has been done with a view to accelerate learning speed. In the medical context, however, generalisation may be more important than learning speed. With the brain stem auditory evoked potential classification task described in this study, the authors found that parameter values that gave fastest learning could result in poor generalisation. In order to achieve maximum generalisation, it was necessary to fine tune the neural net for gain, momentum, batch size, and hidden layer size. Although this maximization could be time consuming, especially with larger training sets, the authors' results suggest that fine tuning parameters can have important clinical consequences, which justifies the time involved. In the authors' case, fine tuning parameters for high generalisation had the additional effect of reducing false negative classifications, with only a small sacrifice in learning speed     

Inbreeding in Swiss Braunvieh and its influence on breeding values predicted from a repeatability animal model.

Inbreeding coefficients were computed for 910,444 animals of the Swiss Braunvieh population. Of the animals born in 1984, 71.5% were inbred with 67.9, 3.4, and .2% having inbreeding coefficients between greater than 0 and 5%, greater than 5 to 10%, and greater than 10%, respectively. The average inbreeding coefficient was 1.14% but, for animals with both parents and at least one grandparent known, it was 1.67%. Breeding values for total milk, fat, and protein yields and for fat and protein percentages were predicted using a repeatability animal model including a regression on the inbreeding coefficient. Phenotypic performance was sizeably depressed for milk yield only (-26 kg/% of inbreeding or 2.4% of the phenotypic standard deviation). Adjusting for inbreeding increased the estimated genetic trend slightly. Inbreeding is only partially accounted for when it is ignored in the construction of the inverse of the numerator relationship matrix. This effect was investigated by comparing predicted breeding values from a model including the complete matrix with predicted breeding values from a model including a matrix constructed with inbreeding ignored. Only .8% of all predicted breeding values were affected by more than +/- 5.5 kg. The maximum difference observed was 55.3 kg. The observed average absolute differences between the breeding values of offspring predicted with the two models increased with inbreeding of parents.     

Means of raising the quality of white titanium enamels

Translated from Steklo i Keramika, No. 10, p. 20, October, 1992.