rafmacalaba commited on
Commit
9af686d
·
verified ·
1 Parent(s): 1ad27bf

Upload holdout_metrics.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. holdout_metrics.json +676 -0
holdout_metrics.json ADDED
@@ -0,0 +1,676 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "real": {
3
+ "n_eval": 9540,
4
+ "overall": {
5
+ "tp": 6956,
6
+ "fp": 2047,
7
+ "fn": 2183,
8
+ "precision": 0.7726313451071865,
9
+ "recall": 0.7611336032388664,
10
+ "f05": 0.7703040907178135,
11
+ "f1": 0.7668393782383419
12
+ },
13
+ "per_key": {
14
+ "producer": {
15
+ "tp": 1474,
16
+ "fp": 568,
17
+ "fn": 752,
18
+ "precision": 0.7218413320274241,
19
+ "recall": 0.6621743036837376,
20
+ "f05": 0.7090629209159129,
21
+ "f1": 0.6907216494845361
22
+ },
23
+ "year": {
24
+ "tp": 1737,
25
+ "fp": 370,
26
+ "fn": 488,
27
+ "precision": 0.8243948742287612,
28
+ "recall": 0.7806741573033708,
29
+ "f05": 0.8152633061109545,
30
+ "f1": 0.8019390581717452
31
+ },
32
+ "geography": {
33
+ "tp": 2159,
34
+ "fp": 751,
35
+ "fn": 652,
36
+ "precision": 0.7419243986254296,
37
+ "recall": 0.768054073283529,
38
+ "f05": 0.7470071275344268,
39
+ "f1": 0.7547631532948785
40
+ },
41
+ "acronym": {
42
+ "tp": 1586,
43
+ "fp": 358,
44
+ "fn": 291,
45
+ "precision": 0.815843621399177,
46
+ "recall": 0.8449653702717101,
47
+ "f05": 0.821506267481612,
48
+ "f1": 0.8301491756084796
49
+ }
50
+ },
51
+ "verbatim": {
52
+ "emitted": 9003,
53
+ "verbatim": 8964,
54
+ "rate": 0.9956681106297901,
55
+ "per_key": {
56
+ "producer": {
57
+ "emitted": 2042,
58
+ "verbatim": 2024
59
+ },
60
+ "year": {
61
+ "emitted": 2107,
62
+ "verbatim": 2107
63
+ },
64
+ "geography": {
65
+ "emitted": 2910,
66
+ "verbatim": 2892
67
+ },
68
+ "acronym": {
69
+ "emitted": 1944,
70
+ "verbatim": 1941
71
+ }
72
+ }
73
+ },
74
+ "usage_impact": {
75
+ "data_type_macro_f1": 0.6985499627733744,
76
+ "usage_action_macro_f1": 0.6339285026503059,
77
+ "impact_label_macro_f1": 0.5977288525673021,
78
+ "per_label": {
79
+ "data_type": {
80
+ "administrative": {
81
+ "tp": 753,
82
+ "fp": 461,
83
+ "fn": 153,
84
+ "precision": 0.6202635914332785,
85
+ "recall": 0.8311258278145696,
86
+ "f1": 0.710377358490566
87
+ },
88
+ "database": {
89
+ "tp": 573,
90
+ "fp": 162,
91
+ "fn": 272,
92
+ "precision": 0.7795918367346939,
93
+ "recall": 0.6781065088757396,
94
+ "f1": 0.7253164556962025
95
+ },
96
+ "geospatial": {
97
+ "tp": 152,
98
+ "fp": 38,
99
+ "fn": 65,
100
+ "precision": 0.8,
101
+ "recall": 0.7004608294930875,
102
+ "f1": 0.7469287469287469
103
+ },
104
+ "indicator": {
105
+ "tp": 174,
106
+ "fp": 77,
107
+ "fn": 137,
108
+ "precision": 0.6932270916334662,
109
+ "recall": 0.5594855305466238,
110
+ "f1": 0.6192170818505338
111
+ },
112
+ "other": {
113
+ "tp": 237,
114
+ "fp": 142,
115
+ "fn": 193,
116
+ "precision": 0.6253298153034301,
117
+ "recall": 0.5511627906976744,
118
+ "f1": 0.5859085290482077
119
+ },
120
+ "report": {
121
+ "tp": 211,
122
+ "fp": 111,
123
+ "fn": 168,
124
+ "precision": 0.65527950310559,
125
+ "recall": 0.5567282321899736,
126
+ "f1": 0.6019971469329529
127
+ },
128
+ "survey": {
129
+ "tp": 1293,
130
+ "fp": 142,
131
+ "fn": 145,
132
+ "precision": 0.9010452961672474,
133
+ "recall": 0.8991655076495132,
134
+ "f1": 0.9001044204664115
135
+ }
136
+ },
137
+ "usage_action": {
138
+ "analyze": {
139
+ "tp": 979,
140
+ "fp": 349,
141
+ "fn": 629,
142
+ "precision": 0.7371987951807228,
143
+ "recall": 0.6088308457711443,
144
+ "f1": 0.6668937329700272
145
+ },
146
+ "curate": {
147
+ "tp": 553,
148
+ "fp": 234,
149
+ "fn": 290,
150
+ "precision": 0.7026683608640406,
151
+ "recall": 0.6559905100830368,
152
+ "f1": 0.6785276073619633
153
+ },
154
+ "inform": {
155
+ "tp": 546,
156
+ "fp": 473,
157
+ "fn": 229,
158
+ "precision": 0.535819430814524,
159
+ "recall": 0.704516129032258,
160
+ "f1": 0.608695652173913
161
+ },
162
+ "source": {
163
+ "tp": 730,
164
+ "fp": 377,
165
+ "fn": 224,
166
+ "precision": 0.6594399277326106,
167
+ "recall": 0.7651991614255765,
168
+ "f1": 0.7083939835031537
169
+ },
170
+ "validate": {
171
+ "tp": 160,
172
+ "fp": 125,
173
+ "fn": 186,
174
+ "precision": 0.5614035087719298,
175
+ "recall": 0.4624277456647399,
176
+ "f1": 0.5071315372424722
177
+ }
178
+ },
179
+ "impact_label": {
180
+ "evidence-base": {
181
+ "tp": 2050,
182
+ "fp": 350,
183
+ "fn": 647,
184
+ "precision": 0.8541666666666666,
185
+ "recall": 0.7601038190582128,
186
+ "f1": 0.8043947420051011
187
+ },
188
+ "none": {
189
+ "tp": 730,
190
+ "fp": 426,
191
+ "fn": 270,
192
+ "precision": 0.6314878892733564,
193
+ "recall": 0.73,
194
+ "f1": 0.6771799628942485
195
+ },
196
+ "policy": {
197
+ "tp": 586,
198
+ "fp": 366,
199
+ "fn": 192,
200
+ "precision": 0.615546218487395,
201
+ "recall": 0.7532133676092545,
202
+ "f1": 0.677456647398844
203
+ },
204
+ "uptake": {
205
+ "tp": 8,
206
+ "fp": 10,
207
+ "fn": 43,
208
+ "precision": 0.4444444444444444,
209
+ "recall": 0.1568627450980392,
210
+ "f1": 0.23188405797101447
211
+ }
212
+ }
213
+ },
214
+ "usage_summary": {
215
+ "n": 4526,
216
+ "mean_sim": 0.5934299230575562,
217
+ "grounded_rate": 0.7551922227132125
218
+ }
219
+ },
220
+ "examples": [
221
+ {
222
+ "mention": "SARA",
223
+ "gold": "{\"acronym\": \"SARA\", \"producer\": \"World Health Organization\"}",
224
+ "pred": "{\"acronym\": \"SARA\", \"producer\": \"USAID\"}",
225
+ "correct": "acronym",
226
+ "verbatim": "producer,acronym"
227
+ },
228
+ {
229
+ "mention": "Encuesta Nacional Industrial Annual",
230
+ "gold": "{\"acronym\": \"ENIA\", \"geography\": \"Chilean\", \"producer\": \"National Statistical Institute\"}",
231
+ "pred": "{\"acronym\": \"ENIA\", \"geography\": \"Chilean\"}",
232
+ "correct": "geography,acronym",
233
+ "verbatim": "geography,acronym"
234
+ },
235
+ {
236
+ "mention": "child protection services data",
237
+ "gold": "{\"year\": \"2018\", \"geography\": \"Jordan\"}",
238
+ "pred": "{\"year\": \"2018\", \"geography\": \"Jordan\"}",
239
+ "correct": "year,geography",
240
+ "verbatim": "year,geography"
241
+ },
242
+ {
243
+ "mention": "GBD data",
244
+ "gold": "{\"geography\": \"Georgia\", \"year\": \"2017\"}",
245
+ "pred": "{\"acronym\": \"GBD\", \"geography\": \"Georgia\", \"year\": \"2017\"}",
246
+ "correct": "year,geography",
247
+ "verbatim": "year,geography,acronym"
248
+ },
249
+ {
250
+ "mention": "Human Capital Index",
251
+ "gold": "{\"producer\": \"World Bank\", \"geography\": \"Djibouti\"}",
252
+ "pred": "{\"producer\": \"World Bank\", \"year\": \"2019\", \"geography\": \"Djibouti\", \"acronym\": \"HDI\"}",
253
+ "correct": "producer,geography",
254
+ "verbatim": "producer,year,geography,acronym"
255
+ },
256
+ {
257
+ "mention": "night-time lights data",
258
+ "gold": "{\"geography\": \"district level\"}",
259
+ "pred": "{\"geography\": \"rural India\"}",
260
+ "correct": "-",
261
+ "verbatim": "geography"
262
+ },
263
+ {
264
+ "mention": "annual production of four major crops",
265
+ "gold": "{\"geography\": \"global\", \"producer\": \"Iizumi and Sakai ( 2020 )\", \"year\": \"2020\"}",
266
+ "pred": "{\"geography\": \"global\", \"year\": \"2020\"}",
267
+ "correct": "year,geography",
268
+ "verbatim": "year,geography"
269
+ },
270
+ {
271
+ "mention": "registry data",
272
+ "gold": "{\"producer\": \"UNHCR\"}",
273
+ "pred": "{\"producer\": \"UNHCR\"}",
274
+ "correct": "producer",
275
+ "verbatim": "producer"
276
+ },
277
+ {
278
+ "mention": "DTM Mali – Septembre 2019",
279
+ "gold": "{\"year\": \"2019\", \"geography\": \"Mali\", \"acronym\": \"DTM\"}",
280
+ "pred": "{\"year\": \"2019\", \"geography\": \"Mali\", \"acronym\": \"DTM\"}",
281
+ "correct": "year,geography,acronym",
282
+ "verbatim": "year,geography,acronym"
283
+ },
284
+ {
285
+ "mention": "natural disaster information",
286
+ "gold": "{\"geography\": \"LAC\"}",
287
+ "pred": "{\"geography\": \"LAC\"}",
288
+ "correct": "geography",
289
+ "verbatim": "geography"
290
+ },
291
+ {
292
+ "mention": "Enquˆete Nationale sur les Conditions de Vie des",
293
+ "gold": "{\"acronym\": \"ECVMA\", \"geography\": \"Niger\"}",
294
+ "pred": "{\"acronym\": \"ECVMA\", \"geography\": \"Niger\"}",
295
+ "correct": "geography,acronym",
296
+ "verbatim": "geography,acronym"
297
+ },
298
+ {
299
+ "mention": "Covid-19 Trade Watch",
300
+ "gold": "{\"producer\": \"World Bank\", \"year\": \"2020\"}",
301
+ "pred": "{\"producer\": \"World Bank\", \"year\": \"2020\"}",
302
+ "correct": "producer,year",
303
+ "verbatim": "producer,year"
304
+ }
305
+ ]
306
+ },
307
+ "synthetic": {
308
+ "n_eval": 2305,
309
+ "overall": {
310
+ "tp": 1005,
311
+ "fp": 155,
312
+ "fn": 165,
313
+ "precision": 0.8663793103448276,
314
+ "recall": 0.8589743589743589,
315
+ "f05": 0.8648881239242684,
316
+ "f1": 0.8626609442060087
317
+ },
318
+ "per_key": {
319
+ "producer": {
320
+ "tp": 252,
321
+ "fp": 23,
322
+ "fn": 35,
323
+ "precision": 0.9163636363636364,
324
+ "recall": 0.8780487804878049,
325
+ "f05": 0.9084354722422495,
326
+ "f1": 0.8967971530249111
327
+ },
328
+ "year": {
329
+ "tp": 259,
330
+ "fp": 30,
331
+ "fn": 46,
332
+ "precision": 0.8961937716262975,
333
+ "recall": 0.8491803278688524,
334
+ "f05": 0.8863791923340176,
335
+ "f1": 0.8720538720538721
336
+ },
337
+ "geography": {
338
+ "tp": 355,
339
+ "fp": 62,
340
+ "fn": 42,
341
+ "precision": 0.8513189448441247,
342
+ "recall": 0.8942065491183879,
343
+ "f05": 0.8595641646489104,
344
+ "f1": 0.8722358722358723
345
+ },
346
+ "acronym": {
347
+ "tp": 139,
348
+ "fp": 40,
349
+ "fn": 42,
350
+ "precision": 0.776536312849162,
351
+ "recall": 0.7679558011049724,
352
+ "f05": 0.7748049052396878,
353
+ "f1": 0.7722222222222223
354
+ }
355
+ },
356
+ "verbatim": {
357
+ "emitted": 1160,
358
+ "verbatim": 1156,
359
+ "rate": 0.996551724137931,
360
+ "per_key": {
361
+ "producer": {
362
+ "emitted": 275,
363
+ "verbatim": 275
364
+ },
365
+ "year": {
366
+ "emitted": 289,
367
+ "verbatim": 289
368
+ },
369
+ "geography": {
370
+ "emitted": 417,
371
+ "verbatim": 413
372
+ },
373
+ "acronym": {
374
+ "emitted": 179,
375
+ "verbatim": 179
376
+ }
377
+ }
378
+ },
379
+ "usage_impact": {
380
+ "data_type_macro_f1": 0.7857483174178359,
381
+ "usage_action_macro_f1": 0.6624851717052487,
382
+ "impact_label_macro_f1": 0.5738134926381705,
383
+ "per_label": {
384
+ "data_type": {
385
+ "administrative": {
386
+ "tp": 247,
387
+ "fp": 82,
388
+ "fn": 31,
389
+ "precision": 0.7507598784194529,
390
+ "recall": 0.8884892086330936,
391
+ "f1": 0.8138385502471169
392
+ },
393
+ "database": {
394
+ "tp": 190,
395
+ "fp": 50,
396
+ "fn": 69,
397
+ "precision": 0.7916666666666666,
398
+ "recall": 0.7335907335907336,
399
+ "f1": 0.7615230460921844
400
+ },
401
+ "geospatial": {
402
+ "tp": 93,
403
+ "fp": 18,
404
+ "fn": 24,
405
+ "precision": 0.8378378378378378,
406
+ "recall": 0.7948717948717948,
407
+ "f1": 0.8157894736842105
408
+ },
409
+ "indicator": {
410
+ "tp": 117,
411
+ "fp": 18,
412
+ "fn": 37,
413
+ "precision": 0.8666666666666667,
414
+ "recall": 0.7597402597402597,
415
+ "f1": 0.8096885813148789
416
+ },
417
+ "other": {
418
+ "tp": 114,
419
+ "fp": 60,
420
+ "fn": 56,
421
+ "precision": 0.6551724137931034,
422
+ "recall": 0.6705882352941176,
423
+ "f1": 0.6627906976744184
424
+ },
425
+ "report": {
426
+ "tp": 30,
427
+ "fp": 9,
428
+ "fn": 15,
429
+ "precision": 0.7692307692307693,
430
+ "recall": 0.6666666666666666,
431
+ "f1": 0.7142857142857142
432
+ },
433
+ "survey": {
434
+ "tp": 564,
435
+ "fp": 45,
436
+ "fn": 50,
437
+ "precision": 0.9261083743842364,
438
+ "recall": 0.9185667752442996,
439
+ "f1": 0.9223221586263287
440
+ }
441
+ },
442
+ "usage_action": {
443
+ "analyze": {
444
+ "tp": 76,
445
+ "fp": 4,
446
+ "fn": 8,
447
+ "precision": 0.95,
448
+ "recall": 0.9047619047619048,
449
+ "f1": 0.9268292682926829
450
+ },
451
+ "causal": {
452
+ "tp": 56,
453
+ "fp": 19,
454
+ "fn": 52,
455
+ "precision": 0.7466666666666667,
456
+ "recall": 0.5185185185185185,
457
+ "f1": 0.6120218579234973
458
+ },
459
+ "curate": {
460
+ "tp": 32,
461
+ "fp": 2,
462
+ "fn": 8,
463
+ "precision": 0.9411764705882353,
464
+ "recall": 0.8,
465
+ "f1": 0.8648648648648648
466
+ },
467
+ "inform": {
468
+ "tp": 138,
469
+ "fp": 48,
470
+ "fn": 51,
471
+ "precision": 0.7419354838709677,
472
+ "recall": 0.7301587301587301,
473
+ "f1": 0.736
474
+ },
475
+ "measure": {
476
+ "tp": 188,
477
+ "fp": 146,
478
+ "fn": 82,
479
+ "precision": 0.562874251497006,
480
+ "recall": 0.6962962962962963,
481
+ "f1": 0.6225165562913907
482
+ },
483
+ "model": {
484
+ "tp": 90,
485
+ "fp": 49,
486
+ "fn": 68,
487
+ "precision": 0.6474820143884892,
488
+ "recall": 0.569620253164557,
489
+ "f1": 0.6060606060606062
490
+ },
491
+ "other": {
492
+ "tp": 25,
493
+ "fp": 21,
494
+ "fn": 36,
495
+ "precision": 0.5434782608695652,
496
+ "recall": 0.4098360655737705,
497
+ "f1": 0.4672897196261682
498
+ },
499
+ "produce": {
500
+ "tp": 57,
501
+ "fp": 29,
502
+ "fn": 61,
503
+ "precision": 0.6627906976744186,
504
+ "recall": 0.4830508474576271,
505
+ "f1": 0.5588235294117647
506
+ },
507
+ "source": {
508
+ "tp": 276,
509
+ "fp": 165,
510
+ "fn": 121,
511
+ "precision": 0.6258503401360545,
512
+ "recall": 0.6952141057934509,
513
+ "f1": 0.6587112171837709
514
+ },
515
+ "validate": {
516
+ "tp": 114,
517
+ "fp": 70,
518
+ "fn": 46,
519
+ "precision": 0.6195652173913043,
520
+ "recall": 0.7125,
521
+ "f1": 0.6627906976744186
522
+ },
523
+ "wrangle": {
524
+ "tp": 24,
525
+ "fp": 8,
526
+ "fn": 28,
527
+ "precision": 0.75,
528
+ "recall": 0.46153846153846156,
529
+ "f1": 0.5714285714285714
530
+ }
531
+ },
532
+ "impact_label": {
533
+ "evidence-base": {
534
+ "tp": 942,
535
+ "fp": 197,
536
+ "fn": 86,
537
+ "precision": 0.8270412642669008,
538
+ "recall": 0.9163424124513618,
539
+ "f1": 0.8694047069681587
540
+ },
541
+ "market-use": {
542
+ "tp": 0,
543
+ "fp": 1,
544
+ "fn": 8,
545
+ "precision": 0.0,
546
+ "recall": 0.0,
547
+ "f1": 0.0
548
+ },
549
+ "none": {
550
+ "tp": 54,
551
+ "fp": 19,
552
+ "fn": 42,
553
+ "precision": 0.7397260273972602,
554
+ "recall": 0.5625,
555
+ "f1": 0.6390532544378699
556
+ },
557
+ "operational-use": {
558
+ "tp": 9,
559
+ "fp": 6,
560
+ "fn": 57,
561
+ "precision": 0.6,
562
+ "recall": 0.13636363636363635,
563
+ "f1": 0.22222222222222218
564
+ },
565
+ "policy": {
566
+ "tp": 254,
567
+ "fp": 82,
568
+ "fn": 110,
569
+ "precision": 0.7559523809523809,
570
+ "recall": 0.6978021978021978,
571
+ "f1": 0.7257142857142858
572
+ },
573
+ "uptake": {
574
+ "tp": 73,
575
+ "fp": 0,
576
+ "fn": 2,
577
+ "precision": 1.0,
578
+ "recall": 0.9733333333333334,
579
+ "f1": 0.9864864864864865
580
+ }
581
+ }
582
+ },
583
+ "usage_summary": {
584
+ "n": 1637,
585
+ "mean_sim": 0.5707868337631226,
586
+ "grounded_rate": 0.6823457544288333
587
+ }
588
+ },
589
+ "examples": [
590
+ {
591
+ "mention": "Women Business and Law",
592
+ "gold": "{\"acronym\": \"WBL\", \"producer\": \"World Bank\"}",
593
+ "pred": "{\"acronym\": \"WBL\", \"producer\": \"World Bank\"}",
594
+ "correct": "producer,acronym",
595
+ "verbatim": "producer,acronym"
596
+ },
597
+ {
598
+ "mention": "Quarterly Public Finance Database",
599
+ "gold": "{\"producer\": \"national sources\"}",
600
+ "pred": "{\"producer\": \"national sources\"}",
601
+ "correct": "producer",
602
+ "verbatim": "producer"
603
+ },
604
+ {
605
+ "mention": "IMF data",
606
+ "gold": "{\"geography\": \"40 low-income countries\", \"producer\": \"IMF\"}",
607
+ "pred": "{\"geography\": \"40 low-income countries\", \"producer\": \"IMF\"}",
608
+ "correct": "producer,geography",
609
+ "verbatim": "producer,geography"
610
+ },
611
+ {
612
+ "mention": "yearly data on firms ’ export transactions",
613
+ "gold": "{\"acronym\": \"ECI\", \"producer\": \"INE\"}",
614
+ "pred": "{\"acronym\": \"ECI\", \"producer\": \"INE\"}",
615
+ "correct": "producer,acronym",
616
+ "verbatim": "producer,acronym"
617
+ },
618
+ {
619
+ "mention": "Inspectorate 2011 data",
620
+ "gold": "{\"geography\": \"Tanzania\", \"year\": \"2011\"}",
621
+ "pred": "{\"geography\": \"Tanzania\", \"year\": \"2011\"}",
622
+ "correct": "year,geography",
623
+ "verbatim": "year,geography"
624
+ },
625
+ {
626
+ "mention": "2018 / 19 NLSS data",
627
+ "gold": "{\"acronym\": \"NLSS\", \"geography\": \"Nigeria\", \"producer\": \"National Bureau of Statistics\"}",
628
+ "pred": "{\"acronym\": \"NLSS\", \"geography\": \"Nigeria\", \"producer\": \"Nigeria ’ s National Bureau of Statistics ( NBS )\"}",
629
+ "correct": "geography,acronym",
630
+ "verbatim": "producer,geography,acronym"
631
+ },
632
+ {
633
+ "mention": "administrative panel data",
634
+ "gold": "{\"geography\": \"Colombia\"}",
635
+ "pred": "{\"geography\": \"Colombia\"}",
636
+ "correct": "geography",
637
+ "verbatim": "geography"
638
+ },
639
+ {
640
+ "mention": "India Human Development Survey ( IHDS )",
641
+ "gold": "{\"acronym\": \"( IHDS )\", \"geography\": \"India\"}",
642
+ "pred": "{\"acronym\": \"IHDS\", \"geography\": \"India\"}",
643
+ "correct": "geography",
644
+ "verbatim": "geography,acronym"
645
+ },
646
+ {
647
+ "mention": "OECD-PISA 2012 database",
648
+ "gold": "{\"producer\": \"OECD\", \"year\": \"2012\"}",
649
+ "pred": "{\"geography\": \"Vietnam\", \"producer\": \"OECD\", \"year\": \"2012\"}",
650
+ "correct": "producer,year",
651
+ "verbatim": "producer,year,geography"
652
+ },
653
+ {
654
+ "mention": "DHS figures on access",
655
+ "gold": "{\"year\": \"1997\"}",
656
+ "pred": "{\"year\": \"1997\"}",
657
+ "correct": "year",
658
+ "verbatim": "year"
659
+ },
660
+ {
661
+ "mention": "DHS surveys",
662
+ "gold": "{\"acronym\": \"DHS\", \"geography\": \"Nigeria\"}",
663
+ "pred": "{\"geography\": \"Nigeria, Cameroon, and Chad\"}",
664
+ "correct": "-",
665
+ "verbatim": "geography"
666
+ },
667
+ {
668
+ "mention": "International Country Risk Guide",
669
+ "gold": "{\"acronym\": \"( ICRG )\"}",
670
+ "pred": "{\"acronym\": \"ICRG\"}",
671
+ "correct": "-",
672
+ "verbatim": "acronym"
673
+ }
674
+ ]
675
+ }
676
+ }