Skip to content

Commit

Permalink
update models/scores
Browse files Browse the repository at this point in the history
  • Loading branch information
kermitt2 committed Dec 3, 2022
1 parent ac6146c commit c5ed7bc
Show file tree
Hide file tree
Showing 4 changed files with 794 additions and 0 deletions.
Binary file not shown.
Binary file not shown.
Original file line number Diff line number Diff line change
@@ -0,0 +1,393 @@
-------------> GROBID failed on 0 PDF

1943 PDF files processed in 1920.476 seconds, 0.988407617086979 seconds per PDF file



Evaluation header 100% │█████████████│ 1943/1943 (0:00:55 / 0:00:00)


Evaluation citation 100% │███████████│ 1943/1943 (0:13:12 / 0:00:00)


Evaluation full text 100% │██████████│ 1943/1943 (0:00:26 / 0:00:00)
Evaluation metrics produced in 874.772 seconds
> :grobid-trainer:jatsEval
======= Header metadata =======

Evaluation on 1943 random PDF files out of 1941 PDF (ratio 1.0).

======= Strict Matching ======= (exact matches)

===== Field-level results =====

label accuracy precision recall f1 support

abstract 82.47 16.69 16.38 16.53 1911
authors 98.61 93.55 93.46 93.51 1941
first_author 99.29 96.75 96.65 96.7 1941
keywords 93.69 61.6 59.86 60.71 1380
title 97.15 86.84 86.57 86.7 1943

all (micro avg.) 94.24 72.1 71.42 71.76 9116
all (macro avg.) 94.24 71.09 70.58 70.83 9116


======== Soft Matching ======== (ignoring punctuation, case and space characters mismatches)

===== Field-level results =====

label accuracy precision recall f1 support

abstract 91.93 62.83 61.64 62.23 1911
authors 98.74 94.17 94.08 94.12 1941
first_author 99.32 96.91 96.81 96.86 1941
keywords 94.93 70.02 68.04 69.02 1380
title 98.62 93.8 93.52 93.66 1943

all (micro avg.) 96.71 84.59 83.8 84.19 9116
all (macro avg.) 96.71 83.55 82.82 83.18 9116


==== Levenshtein Matching ===== (Minimum Levenshtein distance at 0.8)

===== Field-level results =====

label accuracy precision recall f1 support

abstract 97.39 89.44 87.76 88.59 1911
authors 99.23 96.49 96.39 96.44 1941
first_author 99.39 97.22 97.11 97.16 1941
keywords 96.76 82.55 80.22 81.37 1380
title 99.58 98.35 98.04 98.2 1943

all (micro avg.) 98.47 93.51 92.64 93.07 9116
all (macro avg.) 98.47 92.81 91.91 92.35 9116


= Ratcliff/Obershelp Matching = (Minimum Ratcliff/Obershelp similarity at 0.95)

===== Field-level results =====

label accuracy precision recall f1 support

abstract 96.6 85.6 83.99 84.79 1911
authors 99.03 95.51 95.41 95.46 1941
first_author 99.29 96.75 96.65 96.7 1941
keywords 95.95 77.03 74.86 75.93 1380
title 99.32 97.11 96.81 96.96 1943

all (micro avg.) 98.04 91.32 90.47 90.89 9116
all (macro avg.) 98.04 90.4 89.54 89.97 9116

===== Instance-level results =====

Total expected instances: 1943
Total correct instances: 215 (strict)
Total correct instances: 883 (soft)
Total correct instances: 1392 (Levenshtein)
Total correct instances: 1260 (ObservedRatcliffObershelp)

Instance-level recall: 11.07 (strict)
Instance-level recall: 45.45 (soft)
Instance-level recall: 71.64 (Levenshtein)
Instance-level recall: 64.85 (RatcliffObershelp)

======= Citation metadata =======

Evaluation on 1943 random PDF files out of 1941 PDF (ratio 1.0).

======= Strict Matching ======= (exact matches)

===== Field-level results =====

label accuracy precision recall f1 support

authors 97.67 83.65 76 79.64 85778
date 99.17 94.35 83.65 88.68 87067
first_author 98.57 90.12 81.85 85.78 85778
inTitle 96.16 73 71.05 72.01 81007
issue 99.6 88.5 82.98 85.65 16635
page 98.8 95.56 84.83 89.87 80501
title 97.27 80.16 75.04 77.52 80736
volume 99.36 95.63 89.18 92.29 80067

all (micro avg.) 98.32 87.38 80.3 83.69 597569
all (macro avg.) 98.32 87.62 80.57 83.93 597569


======== Soft Matching ======== (ignoring punctuation, case and space characters mismatches)

===== Field-level results =====

label accuracy precision recall f1 support

authors 97.74 84.15 76.44 80.11 85778
date 99.17 94.35 83.65 88.68 87067
first_author 98.6 90.3 82.02 85.96 85778
inTitle 97.85 84.91 82.64 83.76 81007
issue 99.6 88.5 82.98 85.65 16635
page 98.8 95.56 84.83 89.87 80501
title 98.84 91.64 85.78 88.62 80736
volume 99.36 95.63 89.18 92.29 80067

all (micro avg.) 98.74 90.77 83.41 86.93 597569
all (macro avg.) 98.74 90.63 83.44 86.87 597569


==== Levenshtein Matching ===== (Minimum Levenshtein distance at 0.8)

===== Field-level results =====

label accuracy precision recall f1 support

authors 98.54 89.88 81.65 85.57 85778
date 99.17 94.35 83.65 88.68 87067
first_author 98.63 90.52 82.22 86.17 85778
inTitle 98.04 86.21 83.91 85.05 81007
issue 99.6 88.5 82.98 85.65 16635
page 98.8 95.56 84.83 89.87 80501
title 99.16 93.98 87.98 90.88 80736
volume 99.36 95.63 89.18 92.29 80067

all (micro avg.) 98.91 92.12 84.66 88.23 597569
all (macro avg.) 98.91 91.83 84.55 88.02 597569


= Ratcliff/Obershelp Matching = (Minimum Ratcliff/Obershelp similarity at 0.95)

===== Field-level results =====

label accuracy precision recall f1 support

authors 98.09 86.68 78.75 82.52 85778
date 99.17 94.35 83.65 88.68 87067
first_author 98.58 90.14 81.86 85.8 85778
inTitle 97.65 83.5 81.27 82.37 81007
issue 99.6 88.5 82.98 85.65 16635
page 98.8 95.56 84.83 89.87 80501
title 99.1 93.56 87.58 90.47 80736
volume 99.36 95.63 89.18 92.29 80067

all (micro avg.) 98.79 91.16 83.78 87.32 597569
all (macro avg.) 98.79 90.99 83.76 87.21 597569

===== Instance-level results =====

Total expected instances: 90125
Total extracted instances: 85176
Total correct instances: 38757 (strict)
Total correct instances: 50856 (soft)
Total correct instances: 55697 (Levenshtein)
Total correct instances: 52253 (RatcliffObershelp)

Instance-level precision: 45.5 (strict)
Instance-level precision: 59.71 (soft)
Instance-level precision: 65.39 (Levenshtein)
Instance-level precision: 61.35 (RatcliffObershelp)

Instance-level recall: 43 (strict)
Instance-level recall: 56.43 (soft)
Instance-level recall: 61.8 (Levenshtein)
Instance-level recall: 57.98 (RatcliffObershelp)

Instance-level f-score: 44.22 (strict)
Instance-level f-score: 58.02 (soft)
Instance-level f-score: 63.54 (Levenshtein)
Instance-level f-score: 59.62 (RatcliffObershelp)

Matching 1 : 67684

Matching 2 : 4228

Matching 3 : 2051

Matching 4 : 735

Total matches : 74698

======= Citation context resolution =======

Total expected references: 90125 - 46.38 references per article
Total predicted references: 85176 - 43.84 references per article

Total expected citation contexts: 139835 - 71.97 citation contexts per article
Total predicted citation contexts: 116677 - 60.05 citation contexts per article

Total correct predicted citation contexts: 98449 - 50.67 citation contexts per article
Total wrong predicted citation contexts: 18228 (wrong callout matching, callout missing in NLM, or matching with a bib. ref. not aligned with a bib.ref. in NLM)

Precision citation contexts: 84.38
Recall citation contexts: 70.4
fscore citation contexts: 76.76

======= Fulltext structures =======

Evaluation on 1943 random PDF files out of 1941 PDF (ratio 1.0).

======= Strict Matching ======= (exact matches)

===== Field-level results =====

label accuracy precision recall f1 support

figure_title 96.75 32.15 25.43 28.4 7058
reference_citation 59.13 57.48 58.63 58.05 134196
reference_figure 95.02 64.47 63.07 63.76 19330
reference_table 99.11 82.72 83.62 83.17 7327
section_title 94.12 71.58 67.65 69.56 27619
table_title 98.84 57.56 54.84 56.16 3784

all (micro avg.) 90.49 60.23 59.98 60.11 199314
all (macro avg.) 90.49 60.99 58.87 59.85 199314


======== Soft Matching ======== (ignoring punctuation, case and space characters mismatches)

===== Field-level results =====

label accuracy precision recall f1 support

figure_title 98.57 78.83 62.35 69.63 7058
reference_citation 61.84 61.72 62.96 62.34 134196
reference_figure 94.9 65.04 63.62 64.32 19330
reference_table 99.08 82.87 83.77 83.32 7327
section_title 94.8 76.13 71.95 73.98 27619
table_title 99.44 81.78 77.91 79.79 3784

all (micro avg.) 91.44 65.57 65.3 65.43 199314
all (macro avg.) 91.44 74.39 70.43 72.23 199314


************************************************************************************
COUNTER: org.grobid.core.engines.counters.TableRejectionCounters
************************************************************************************
------------------------------------------------------------------------------------
CANNOT_PARSE_LABEL_TO_INT: 140
CONTENT_SIZE_TOO_SMALL: 78
CONTENT_WIDTH_TOO_SMALL: 20
EMPTY_LABEL_OR_HEADER_OR_CONTENT: 1991
HEADER_NOT_STARTS_WITH_TABLE_WORD: 148
HEADER_NOT_CONSECUTIVE: 983
HEADER_AND_CONTENT_DIFFERENT_PAGES: 11
HEADER_AND_CONTENT_INTERSECT: 564
FEW_TOKENS_IN_HEADER: 1
====================================================================================

************************************************************************************
COUNTER: org.grobid.core.engines.counters.ReferenceMarkerMatcherCounters
************************************************************************************
------------------------------------------------------------------------------------
UNMATCHED_REF_MARKERS: 9998
MATCHED_REF_MARKERS_AFTER_POST_FILTERING: 3270
STYLE_AUTHORS: 37197
STYLE_NUMBERED: 52094
MANY_CANDIDATES: 4776
MANY_CANDIDATES_AFTER_POST_FILTERING: 604
NO_CANDIDATES: 18385
INPUT_REF_STRINGS_CNT: 91415
MATCHED_REF_MARKERS: 116677
NO_CANDIDATES_AFTER_POST_FILTERING: 500
STYLE_OTHER: 2124
====================================================================================

************************************************************************************
COUNTER: org.grobid.core.engines.counters.FigureCounters
************************************************************************************
------------------------------------------------------------------------------------
SKIPPED_BAD_STANDALONE_FIGURES: 659
SKIPPED_DUE_TO_MISMATCH_OF_CAPTIONS_AND_VECTOR_AND_BITMAP_GRAPHICS: 3
SKIPPED_SMALL_STANDALONE_FIGURES: 526
SKIPPED_BIG_STANDALONE_FIGURES: 133
====================================================================================

************************************************************************************
COUNTER: org.grobid.core.engines.label.TaggingLabelImpl
************************************************************************************
------------------------------------------------------------------------------------
HEADER_DOCTYPE: 2897
CITATION_TITLE: 81615
HEADER_DATE: 1131
HEADER_KEYWORD: 1429
NAME-HEADER_MIDDLENAME: 5839
TABLE_FIGDESC: 4343
NAME-HEADER_SURNAME: 13903
NAME-CITATION_OTHER: 437103
CITATION_BOOKTITLE: 7123
HEADER_FUNDING: 148
HEADER_ADDRESS: 6017
HEADER_AFFILIATION: 6186
CITATION_NOTE: 2842
FULLTEXT_CITATION_MARKER: 181115
TABLE_NOTE: 2999
HEADER_EMAIL: 2210
FULLTEXT_TABLE_MARKER: 14699
CITATION_WEB: 1375
HEADER_GROUP: 4
TABLE_LABEL: 3399
FULLTEXT_SECTION: 55218
NAME-HEADER_FORENAME: 14116
DATE_YEAR: 86760
TABLE_CONTENT: 5357
CITATION_COLLABORATION: 42
CITATION_ISSUE: 17151
HEADER_MEETING: 24
HEADER_EDITOR: 114
CITATION_SERIES: 224
CITATION_JOURNAL: 77697
NAME-CITATION_SURNAME: 330684
TABLE_FIGURE_HEAD: 4837
FULLTEXT_EQUATION_MARKER: 1651
CITATION_OTHER: 450295
FULLTEXT_FIGURE_MARKER: 37742
HEADER_TITLE: 2041
CITATION_TECH: 383
FIGURE_CONTENT: 3283
FIGURE_LABEL: 5990
FULLTEXT_EQUATION_LABEL: 1962
HEADER_OTHER: 10807
FULLTEXT_EQUATION: 4418
TABLE_OTHER: 1
CITATION_DATE: 86066
CITATION_AUTHOR: 86094
FULLTEXT_FIGURE: 14313
FULLTEXT_TABLE: 10073
CITATION_EDITOR: 2699
FULLTEXT_OTHER: 509
HEADER_SUBMISSION: 1207
NAME-HEADER_OTHER: 17369
FIGURE_FIGDESC: 7505
NAME-HEADER_SUFFIX: 20
HEADER_AVAILABILITY: 13
CITATION_VOLUME: 76292
CITATION_LOCATION: 7896
NAME-CITATION_SUFFIX: 394
NAME-HEADER_TITLE: 735
DATE_MONTH: 3107
HEADER_WEB: 344
HEADER_ABSTRACT: 2305
CITATION_INSTITUTION: 1685
HEADER_REFERENCE: 3047
CITATION_PAGES: 80522
HEADER_AUTHOR: 4272
NAME-HEADER_MARKER: 8104
DATE_OTHER: 4721
NAME-CITATION_FORENAME: 319284
CITATION_PUBLISHER: 6061
HEADER_PUBNUM: 1730
NAME-CITATION_MIDDLENAME: 66214
CITATION_PUBNUM: 10886
HEADER_COPYRIGHT: 2379
FULLTEXT_PARAGRAPH: 381002
FIGURE_FIGURE_HEAD: 9715
DATE_DAY: 2836
====================================================================================

************************************************************************************
COUNTER: FigureCounters
************************************************************************************
------------------------------------------------------------------------------------
STANDALONE_FIGURES: 491
ASSIGNED_GRAPHICS_TO_FIGURES: 3777
====================================================================================
====================================================================================
Loading

0 comments on commit c5ed7bc

Please sign in to comment.