Laboratory of Structural Methods of Data Analysis in Predictive
Modeling Moscow Institute of Physics and Technology
Approximation problem for factorized data
We consider samples with factorial design of experiments (full or incomplete). Universal approximation methods don’t take into account peculiarities of such samples. We develop structural approximation method which is based on special function class and regularization. Optimal solution in this class can be found efficiently.

Авторы: Belyaev Mikhail

Дата: 17 ноября 2014

Статус: опубликована

Журнал: Artificial intelligence and decision theory

Выпуск: 3

Год: 2013

Google scholar:

Направления исследований

Structural optimization

Statistical methods

Дополнительные материалы