search for: wbtwb

...so I compared to approaches: - Load 'WB' using 'read.table' (put it in WB.tmp) and then to the simple matrix multiplication WB.tmp%*%t(WB.tmp) - Scan each row of WB and do the cross products 'sum(WB.i*WB.i)' and 'sum(WB.i*WB.j)', which proper arrangement leads to WBtWB. Comparing these two matrices, I get the very similar values, however when I tried their inverse, WBtWB leads to a singular system. I've tried different tests and my conclusion is that my precision problem is related to cross products 'sum(WB.i*WB.i)' and 'sum(WB.i*WB.j)'. Do...