100 likes | 223 Vues
This section explores the concept of projecting data points onto vectors within multivariate data analysis. If ( X' ) represents an ( n times p ) data matrix and ( a ) is a ( p times 1 ) vector, the projection ( Y' = X'a ) yields coordinates of each observation along the vector ( a ). We illustrate this with an example where points such as ( x_1 = (1, 2)', x_2 = (2, 1)', x_3 = (-1, 1)' ) are projected onto the vector ( a = (3, -4)/5 ). The resulting coordinates describe their positions relative to the direction of the vector.
E N D
Aside: projections onto vectors • If X is an np data matrix and a a p1 vector then Y= Xais the projection of X onto a • Values of Y give the coordinates of each observation along the vector a • Example: x1=(1, 2), x2=(2, 1), x3=(1,1) a=(3, 4)/5, • So Multivariate Data Analysis
Aside: projections onto vectors x1 2 x3 x2 (0,0) 1 Multivariate Data Analysis
Aside: projections onto vectors x1 2 x3 x2 (0,0) 1 a=(3, 4)/5 Multivariate Data Analysis
Aside: projections onto vectors x1 2 x3 x2 y1 (0,0) 1 a=(3, 4)/5 Multivariate Data Analysis
Aside: projections onto vectors x1=(1, 2) 2 x3 x2 y1 = (3142)/5 = 1 (0,0) 1 a=(3, 4)/5 Multivariate Data Analysis
Aside: projections onto vectors x1 2 x3 x2 y1 = 1 (0,0) 1 y2 = 2/5 a=(3, 4)/5 Multivariate Data Analysis
Aside: projections onto vectors x1 2 y3 = 7/5 x3 x2 y1 = 1 (0,0) 1 y2 = 2/5 a=(3, 4)/5 Multivariate Data Analysis
Aside: projections onto vectors 2 y3 = 7/5 y1 = 1 (0,0) 1 y2 = 2/5 a=(3, 4)/5 Multivariate Data Analysis
Aside: projections onto vectors 2 (0,0) 1 a=(3, 4)/5 Multivariate Data Analysis