zbMATH — the first resource for mathematics

Further aspects of the theory of multiple regression. (English) Zbl 0018.15803
This is a continuance of a previous paper of the author [Proc. Camb. Philos. Soc. 30, 327–340 (1934; Zbl 0010.07102)] in which problems in the theory and application of linear multiple regression are dealt with by vector methods. In particular, results of Hotelling on his ”most predictable criterion” and on the resolution into principal components, and those of Fisher on linear discriminant functions are deduced and discussed in vector form. The \(\chi^2\) approximation for the \(\Lambda\) test (loc. cit.) is further discussed.
Reviewer: C. C. Craig

62J05 Linear regression; mixed models