作者: Catalina Garcia , Roman Salmeron Gomez , Jose Garcia Perez , Maria Del Mar Lopez Martin ,
DOI: 10.17485/IJST/2017/V10I13/98678
关键词: Multicollinearity 、 Mathematics 、 Correlation 、 Relation (database) 、 Value (mathematics) 、 Econometrics 、 Variables 、 Selection (genetic algorithm) 、 Ridge (differential geometry) 、 Statistics 、 Collinearity
摘要: Objectives: To analyze the relation of ridge and raise factors with squared coefficient correlation present, from this relation, a methodology to select adequate value factors. Methods/Statistical Analysis: Two independent variables have been simulated given between 0.95 0.999 gap 0.001 obtain dependent variable. Then, it has selected factor that leads values VIF lesser than 10 for every correlation. Findings: Apart propose way factor, is concluded linear while potential. Application/Improvements: The procedure presented in paper allows adequately selecting mitigates collinearity. This contribution can be applied many different fields where collinearity present.