[Application of analysis method of orthogonal polynomial regression to sports training]

During studying the dynamic changing rules of athlete's physiological workload in training by means of the polynomial regression method in recent years, we found: (1) If the number of the polynomial term or the order is relatively large (the order number M > 6), the equation's calculating amounts increase greatly and it is easy to have the phenomenon of accession to an "ill state" of the regression matrix or the matrix of the proper equation coefficients as well; (2) Both the accuracy for the fitted results and the stability decrease if there are strong correlation between the different terms in the polynomial (between the different odd power terms or even power terms). In order to overcome these defects we employed the orthogonal polynomial regression method and achieved a better effect.