A direct method for building sparse kernel learning algorithms. The Journal of Machine Learning Research 7, 624 (2006) A Derivation of Equation 3 Denote by Gi· the ith row of G. Wewrite ∑ zicp(xi) = ∑ ( zi ( ∑ pjGij) 2 ∑ − 2ri pjGij + r 2 ) i i i = ∑ i j (0)

by M Wu, B Schölkopf, G Bakır
Venue:i·Gi· p − 2 } i {{ } i QL(z) = p T QL(z)p − 2w T L (z)p + CL, ) ziriGi· } {{ } wT L (z) ) p + ∑ Using similar manipulations we obtain ∑ (1 − zi)cq(xi) =q T QR(z)q − 2w T R (z)q + CR. i