multi class SVM in R - weights extraction

Question

My question is about my 28 feature dataset with 6 classes

1. When running t(w %*% t(as.matrix(testset[,2:29]))) - svm.model$rho for the 6 class problem, I get a matrix showing the testing dataset and the classes minus 1 (5 instead of 6), same as the sample table below. How can I determine the class from this table?

               V1          V2          V3          V4          V5
   1   3191    172.58302   -1527.0875  -1301.15106 -767.5058   3451.69551
   2   2159    1199.02091  -1782.2845  -1362.27901 -2257.9054  4132.92307
   3   4295    1557.35203  -2374.2095  -1581.37368 -3601.2268  5963.68623
   4   3843    433.74900   -1091.0102  -1296.25825 -1285.6248  3217.86395
   5   4448    1184.22539  -1515.8411  -1708.36731 -2612.3417  4621.73283
   

2. How can the above be generated from where the

   • Bias is a matrix (15*1) ... (15 values)
   • Weights is a matrix (5*28)
   • And any new point (x) is 28 features