|
ИСТИНА |
Войти в систему Регистрация |
Интеллектуальная Система Тематического Исследования НАукометрических данных |
||
The fuzzy classification of molecular graphs and fast rejection rules in «structure – property» problemтезисы доклада
- Авторы: Prokhorov E.I., Ponomareva L.A., Permyakov E.A., Kumskov M.I.
- Сборник: Proc. 10th Int. Conf. Pattern Recognition And Image Analysis: New Information Technologies
- Том: 2
- Тезисы
- Год издания: 2010
- Место издания: St. Petersburg
- Первая страница: 217
- Последняя страница: 220
- Аннотация: The paper proposes a new approach to the analysis of the matrix "molecule - descriptor" the problem in «structure – property» based on fuzzy cluster structure of the training sample. Describes methods of constructing the "fast" rejection rules and search for ejections in the training set. For this, we propose special space of descriptors, which are simple for a calculating. We consider the optimization of the classifying function of the parameters of fuzzy clustering. We have constructed predictive models with high quality of prediction based on the proposed approach. The paper includes the comparison of models, which indicate the effectiveness of these methods.
- Добавил в систему: Кумсков Михаил Иванович