How is the ARROW derived?

2 replies · 3K views · last reply
tuanle007
Messages
35
Reaction score
0
can someone explain to me how did it derived to the ARROW...
I am not quite sure how that happened.thanks
 
Attachments
  • Handwritten-style derivation of a bivariate Gaussian conditional density and E[Y | X = x] = rho y, with completing the square step and integrals over x and y.
    untitled.PNG
    31.7 KB · Views: 502
You have
[tex]e^{-\frac{x^2+ y^2- 2\rho xy}{2\sigma(1-\rho)^2}}[/itex]<br /> They complete the square:<br /> [tex]x^2- 2\rho xy+ \rho^2y^2- \rho^2y^2+ y^2= (x- \rho y)^2+ (y^2- \rho^2 y^2)[/tex]<br /> so the exponential becomes<br /> [tex]e^{-\frac{(x-\rho y)^2- (y^2- \rho^2 y^2)}{2\sigma(1-\rho)^2}}[/itex]<br /> Then separate the exponentials:<br /> [tex]e^{-\frac{(x- \rho y)^2}{2\sigma(1-\rho)^2}}e^{-\frac{(y^2- \rho^2 y^2}{2\sigma(1-\rho)^2}}[/tex]<br /> and finally factor out y<sup>2</sup> in the last exponential<br /> [tex]e^{-\frac{(x- \rho y)^2}{2\sigma(1-\rho)^2}}e^{-\frac{y^2(1- \rho^2}{2\sigma(1-\rho)^2}}[/tex][/tex][/tex]