INTUITIVE METHOD OF FACTORS SIGNIFICANCE RANKING

UDC 519.242

A.A.Moiseev

Effectiveness of fractional replicas application increases at non – fiction factors multiplication. It depends also on linear and non. – linear effects and factor experiments performance scheme in the presence of factors interactions. In this situation a priory information is desirable and even necessary. Developed here intuitive method of factors ranking in linear regression model based on regular plans and posterior rejection of least loaded factors. Its main advantage along with realization simplicity is only non fictive factors application. Base disadvantage along with intuitivism is swift effectiveness decrease at factors number rise.

Keywords: factor experiment, linear regression model, regular plan, relative load, ranking, non fictive factors, generating functions.

Full text:
Moiseev_1_17_2.pdf