Please answer the following question. Thank you.

Two-Class LDA in 2 Dimensions: Computation 2 points possible (graded) As above, consider the setting of two-class LDA in RP; that is, your data consist of two classes modeled as two Gaussians with the same covariance matrix. But for this problem, your data is in IR, and the estimated parameters are as follows: XC = 0 ~ N( Mo; E) where Mo = (o XC = 1 ~ N( M1, E) where M1 = ( ) The covariance matrix > has eigenvalues o and 1 with corresponding eigenvectors V,2 and V1. In diagonalized form, > is E = V 02 where Vo2 = V1 V2 1 VI =The multivariate Gaussian distribution for classes 0 and 1 both have - - - - T _ L variance 0'2 In the direction of v02 ( 1 1) and variance 1 in the direction of v? = i( 1 1 ) 1. If 0'2 = 1, what is the normal vector n of the decision boundary? (No computation should be needed.) 1 (Enter as a vector, e.g. type [1,2] for the vector (2 ). All scalar multiples are accepted.) For 0'2 = 1, n = 2. For the general case (when 0'2 is not necesssarilyr 1), find the normal W of the decision boundary in terms of 0'2. (All scalar multiples are accepted.)