Clustered by feature linear regression on data with real value

Бесплатный доступ

The authors offer a linear regression modification, based on preset clustering by char. value. Approbation of this method with some real-valued features databases showed highly better results, in comparison with the classical linear regression. The fuzzy sets method was applied to generate discrete features from the real ones.

Data mining, linear regression, clustering, fuzzy sets

Короткий адрес: https://sciup.org/148176873

IDR: 148176873

Статья научная