D13-1125 the training data ( in 10-fold crossvalidation for SVMs ) . Training and test
D13-1077 experiments are performed via 10-fold crossvalidation on documents . We use gold standard
D09-1047 both parameters through 10-fold crossvalidation on the training set . For every
D08-1036 accuracy of the resulting tagging the crossvalidation accuracy . Finally , following
D12-1081 dataset , we performed 5-fold crossvalidation for both classification tasks
D08-1001 we generally performed 3-fold crossvalidation for all performance measures
C00-1051 tests , we conducted five-fold crossvalidation on the whole sentence set . 2In
D09-1047 training parameters through 10-fold crossvalidation on the training set . By treating
C94-1023 can not be completely removed by crossvalidation process ( i.e. , pruning process
D09-1083 experiments using the same 10-fold crossvalidation setting described in Sec . 4.1.3
D08-1054 and test ) and conducted 3-fold crossvalidation experiments . In each trial ,
D12-1066 are still given by performing crossvalidation averaging over 10 test sets for
C04-1058 piped ensemble that incorporates crossvalidation style n-fold partition sets generated
A00-1031 re-examined , since we use a ten-fold crossvalidation and averaging of results while
D09-1095 learning algorithms based on 10-fold crossvalidation on the training data . We have
D09-1158 is shown in Table 4 ( by 2-fold crossvalidation with random ordering ) as a reference
D09-1158 is shown in Table 3 ( by 2-fold crossvalidation with random ordering ) as a reference
D13-1010 Bayes classification in a 5-fold crossvalidation setup . We vary A for the SAGE
C04-1136 of ranking methods . Since the crossvalidation would have to be based on random
D08-1001 334 testing examples . From the crossvalidation runs , a 95 % - confidence interval
hide detail