1. Reported is the test error (TeVErr), for the model selected such
that the cross validation error (see average over 10 folds as reported
in ip_don_all_results.txt) is smallest (independent on each of the
five realizations of the data set.).

run:1 d1:3 d2:1 C:1.50e+00 l:3 => TeVErr:6.6955e-02 #svs=396
run:2 d1:2 d2:1 C:1.00e+00 l:4 => TeVErr:6.0417e-02 #svs=371
run:3 d1:3 d2:1 C:1.00e+00 l:4 => TeVErr:6.8670e-02 #svs=440
run:4 d1:2 d2:2 C:4.00e+00 l:5 => TeVErr:7.1882e-02 #svs=668
run:5 d1:3 d2:1 C:1.00e+00 l:6 => TeVErr:4.4681e-02 #svs=492
run:6 d1:3 d2:1 C:2.00e+00 l:4 => TeVErr:5.8427e-02 #svs=440
run:7 d1:2 d2:2 C:1.00e+00 l:3 => TeVErr:6.8323e-02 #svs=648
run:8 d1:3 d2:1 C:2.00e+00 l:5 => TeVErr:6.4449e-02 #svs=439
run:9 d1:3 d2:1 C:2.00e+00 l:5 => TeVErr:4.6809e-02 #svs=472
run:10 d1:3 d2:1 C:4.00e+00 l:6 => TeVErr:6.7416e-02 #svs=491

TeVErr average+/-std. deviation: 6.18 +/- 0.94 %