Partial Least Squares Improvement and Research Principal Component Regression Extraction Methods

摘要

Partial least squares algorithm as a new type of multivariate data analysis methods, the number of cross-validation to determine the effect of poor primary component. In this paper, cross-validation, based on the interpretation of the degree of integration of the main components of the independent variables and the dependent variable, the importance of the variables as a test indicators, the data on the efficacy indexes were calculated, compared with more conventional partial least squares consistent with the theoretical value, indicating that this algorithm can solve the partial least squares method to determine the main issues to score.