Skip to main content

Advertisement

Table 6 Precision, recall, and F-measure results for labels and spans

From: An annotation and modeling schema for prescription regimens

Tag # train Balanced P & R Precision Bias Recall Bias
P R F P R F P R F
Dispense 46 1.000 0.333 0.500 1.000 0.364 0.222 1.000 0.333 0.500
Dispense_unit 73 0.857 0.316 0.462 1.000 0.348 0.211 1.000 0.316 0.480
Doseamount 221 0.900 0.614 0.730 1.000 0.275 0.159 1.000 0.500 0.667
Doseamount_unit 202 0.921 0.761 0.833 1.000 0.386 0.239 1.000 0.630 0.773
Doseform 533 0.975 0.826 0.895 1.000 0.769 0.625 0.981 0.736 0.841
Duration 205 1.000 0.754 0.860 1.000 0.659 0.492 1.000 0.590 0.742
Duration_unit 193 1.000 0.772 0.871 1.000 0.705 0.544 1.000 0.579 0.733
FREQ 1329 0.978 0.735 0.840 0.990 0.683 0.522 0.977 0.700 0.816
Indication 319 0.879 0.296 0.443 1.000 0.310 0.184 0.852 0.235 0.368
Instruction 4075 0.602 0.892 0.719 1.000 0.068 0.035 0.510 0.928 0.658
Medication 145 1.000 0.103 0.188 1.000 0.000 0.000 1.000 0.103 0.188
PRN 232 1.000 0.855 0.922 1.000 0.804 0.673 1.000 0.782 0.878
Refill 70 1.000 0.000 0.000 1.000 0.000 0.000 1.000 0.000 0.000
Route 591 0.947 0.423 0.584 1.000 0.408 0.256 0.884 0.363 0.515
Strength 33 1.000 0.000 0.000 1.000 0.000 0.000 1.000 0.000 0.000
Strength_unit 34 1.000 0.000 0.000 1.000 0.000 0.000 1.000 0.000 0.000
Sub_status 112 0.929 0.500 0.650 1.000 0.000 0.000 1.000 0.462 0.632
Take 774 0.952 0.855 0.901 0.993 0.726 0.573 0.962 0.761 0.850
Timing 1359 0.979 0.634 0.770 1.000 0.506 0.339 0.908 0.564 0.696
Overall 10,546 0.743 0.753 0.748 0.996 0.407 0.256 0.651 0.726 0.687
  1. For each tag, the highest performing f-measure is presented in boldface