Research on Several Problems in Partial Least Squares Regression Analysis

ABSTRACT

Purpose: preliminary discussion on model prediction precision in the partial least squares regression analysis method; Method: introduce current development conditions of partial least squares regression analysis, analyze problems of traditional regression analysis method such as multiple linear regression analysis, introduce the mathematic principle and modeling method of the partial least squares regression analysis method, and conduct detailed analysis on the partial least squares regression analysis modeling and prediction by using the classical Linnerud data. Result: The partial least squares regression analysis has the basic features of the multiple linear regression analysis and principal component analysis, can precisely predict multiple data and establish a precise mathematical model; Conclusion: The partial least squares regression analysis can provide precise mathematical model and can reserve the explaining variants remarkably associated to explained variants to most extent, so it is feasible to some extent and can meet the general requirements of engineering, economy, biology and medical statistical analysis.

Keywords:

Partial least squares regression analysis, model prediction, multiple linear regression analysis, principal component analysis, medical statistics.