Аннотация
We propose a neural net solution for the recognition of the domain types of proteins, which is a hard and important problem in biology. We have found that using a clever preprocessing technique relatively small neural networks perform surprisingly well. The performances of the neural nets were measured by cross-validation and Hoeffding's inequality was utilized for the estimation of a confidence interval of the estimates.
Пользователи данного ресурса
Пожалуйста,
войдите в систему, чтобы принять участие в дискуссии (добавить собственные рецензию, или комментарий)