I have done with labeling my dataset. I am very happy of that.
(I had pre-annotated data with regex, I correct the annotation by ner.manual for each label, then I merged the labels, then I used ner.batch-train to make a spaCy model) (labels are DATA,PARA,TIME AND so…)
me, and you will cease to wonder that the fifth year has now gone by since I took up Mars,12 PLAN although the year 1603 DATE was nearly all given over to optical investigations.There will arise subtle geometers such as Vieta who will think it something great to show up the contrived nature13 of this method.For in this matter, Vieta did object to Ptolemy, Copernicus, NAME and Regiomontanus NAME ^14.Let them therefore go forth themselves and solve the figure geometrically, and they will be to me great Apolloes ^15.For me it is enough to draw four or five conclusions from a single argument (which includes four observations and two hypotheses); that is, in getting from the labyrinth back to the highway, to show, instead of a geometrical light, a contrived thread,16 which nonetheless will lead you to the exit.If this method is difficult to grasp, the subject is much more difficult to investigate with no method at all.There now follows an example of this instruction based upon the four proposed observations.Because of precession, all positions are reduced to [the time of] the first observation.Here, the apparent longitude was 25° 43’ Virgo, the mean longitude 6s 0° 47’ 40“, and the annual motion of the fixed stars is 51 seconds, as Brahe has demonstrated in the Progymnasmata.Therefore, from 1587 March 6 DATE to 1591 June 8 DATE is 4 years 3 months, to which corresponds a motion of precession of 3’ 37“.Therefore, we must set the apparent position in 1591 DATE at 26° 39’ 23“ Sagittarius, and the mean longitude at 9s 5° 40’ 18“.Similarly, from 1587 March 6 DATE to 1593 August 25 DATE is 6 years 5½ months, to which corresponds a motion of precession of 5’ 30“.And PARA so Mars PLAN is to be placed at 12° 10’ 30“ Pisces, with mean longitude 11s 9° 49’ 34“.Finally, from 1587 March 6 DATE to 1595 October 31 DATE is nearly 8 years 7 months, to which corresponds a motion of 7’ 18“.And so Mars PLAN is to be placed at 17° 24
I have kind of general question,
can you let me know how can I use named entity recognition to build features for sentences?
I also want to know, how can I improve the result of my ner. here is the result: (for instance the result of label data has .44 accuracy,since I consider different type of enitity as data like
“in 1593”
“1587 March 6” or
BEFORE 0.007
Correct 23
Incorrect 3318
Entities 2778
Unknown 0
# LOSS RIGHT WRONG ENTS SKIP ACCURACY
01 103.815 498 164 574 0 0.752
02 98.911 522 135 593 0 0.795
03 98.409 511 131 567 0 0.796
04 97.138 530 117 591 0 0.819
05 97.348 535 119 603 0 0.818
06 97.865 543 94 594 0 0.852
07 95.676 537 101 589 0 0.842
08 96.136 539 97 589 0 0.847
09 96.591 541 94 590 0 0.852
10 95.908 544 95 597 0 0.851
11 96.440 539 92 584 0 0.854
12 97.637 542 100 598 0 0.844
13 95.098 539 96 588 0 0.849
14 95.213 544 91 593 0 0.857
15 95.818 540 97 591 0 0.848
16 97.242 540 89 583 0 0.859
17 96.328 542 83 581 0 0.867
18 95.629 540 88 582 0 0.860
19 97.280 537 99 587 0 0.844
20 96.092 538 97 587 0 0.847
21 96.273 536 97 583 0 0.847
22 96.243 539 102 594 0 0.841
23 96.033 538 96 586 0 0.849
24 96.690 538 94 584 0 0.851
25 95.916 536 94 580 0 0.851
26 96.435 537 93 581 0 0.852
27 95.782 539 90 582 0 0.857
28 96.349 540 94 588 0 0.852
29 95.484 540 92 586 0 0.854
30 96.627 537 95 583 0 0.850
31 96.031 539 97 589 0 0.847
32 95.907 542 95 593 0 0.851
33 96.753 538 95 585 0 0.850
34 95.288 540 99 593 0 0.845
35 96.412 541 97 593 0 0.848
36 97.771 541 92 588 0 0.855
37 95.581 543 89 589 0 0.859
38 96.557 542 91 589 0 0.856
39 95.765 543 95 595 0 0.851
40 97.480 543 89 589 0 0.859
41 97.037 545 90 594 0 0.858
42 95.491 544 87 589 0 0.862
43 95.915 547 87 595 0 0.863
44 96.625 545 93 597 0 0.854
45 96.069 544 88 590 0 0.861
46 96.041 545 90 594 0 0.858
47 96.476 543 94 594 0 0.852
48 96.780 543 93 593 0 0.854
49 95.502 541 93 589 0 0.853
50 95.726 539 95 587 0 0.850
Correct 542
Incorrect 83
Baseline 0.007
Accuracy 0.867
my last question regarding the print command, still it does not work ,can you help me?