id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
bjmsr-311	Bidabad, Bijan	New Algorithms for L1 Norm Regression	2019	18	.pdf	application/pdf	10856	626	72	Substitute (31) in (17), n n S = Σ │yi-ya-ß2 a(x2i-x2 a)│ = Σ │yi-yb-ß2 a(x2i-x2a)│ i=1 i=1 or, n n S = Σ │yi a-ß2 ax2i a│ = Σ │yi b-ß2 ax2i b│ (32) i=1 i=1 Using (30) the first sum in (32) is Sa * and the second sum is Sb evaluated at ß2=ß2 a. Let, n S = Σ │yi - ß2x2i - ß3x3i│ (38) i=1 S can be written as, n n S = Σ │x2i││yi/x2i - ß2 - ß3x3i/x2i│ = Σ │x2i││yi s1 - ß2 - ß3x3i s1│ (39)	cache/bjmsr-311.pdf	txt/bjmsr-311.txt