As stated from EPO, PATSTAT datasets contain patent data for over 100 patent authorities.
What is unknown many times is the quality of such data.
Here we want to peek how backard citations (those appearing on front page of patent application) change across patent authorities and whether or not the patent is the regional phase of a PCT application.
This latter question arises since in such cases you may find citation in espacenet to the document that are not assigned in Patstat (FI EP1659858 has 31 citations in espacenet but none in patstat) reflecting what already existed in REFI (see also citation of WO2005004592) so we want to know at what extent.
So for Patent authorities with more than 50K applications I extracted the number of distinct applications, distinguishing them among being or not regional phase of a WO application (WO vs NOT WO).
Then calculated Citation ratio (# of applications with bw citations / total applications), number of non regional and regional phase application with backward citations (NOT WO CIT / WO CIT) ratio citations / applications for both cases and, eventually it ratio of the last two figures (this indicator if is very small says regional phase application have proportionally less backward citations than other applications)
SOME CONCLUSIONS:
from above data we may say:
1 Coverage of backard citations in national offices is very poor (apart from USPTO) while regional offices (EPO, International Bureau) perform well
2 In most cases backward citations should be enriched with WO equivalent citations (WO/ NON WO ratio < < 1)
thanks to Lorenzo Cassi for rising the issue
What is unknown many times is the quality of such data.
Here we want to peek how backard citations (those appearing on front page of patent application) change across patent authorities and whether or not the patent is the regional phase of a PCT application.
This latter question arises since in such cases you may find citation in espacenet to the document that are not assigned in Patstat (FI EP1659858 has 31 citations in espacenet but none in patstat) reflecting what already existed in REFI (see also citation of WO2005004592) so we want to know at what extent.
So for Patent authorities with more than 50K applications I extracted the number of distinct applications, distinguishing them among being or not regional phase of a WO application (WO vs NOT WO).
Then calculated Citation ratio (# of applications with bw citations / total applications), number of non regional and regional phase application with backward citations (NOT WO CIT / WO CIT) ratio citations / applications for both cases and, eventually it ratio of the last two figures (this indicator if is very small says regional phase application have proportionally less backward citations than other applications)
APPLN AUTH
|
distinct appln_ids
|
NOT WO
|
WO
|
CIT RATIO
|
NOT WO CIT
|
WO CIT
|
WO cit%
|
non wo cit%
|
WO/ NON WO ratio
|
AR
|
105 346
|
104 908
|
438
|
0,00%
|
0
|
0
|
0,00%
|
0,00%
|
|
AT
|
1 139 776
|
984 831
|
154 945
|
0,69%
|
7 814
|
3
|
0,00%
|
0,79%
|
0,0024
|
AU
|
1 626 460
|
941 509
|
684 951
|
16,47%
|
155 131
|
112 720
|
16,46%
|
16,48%
|
0,9988
|
BE
|
648 744
|
648 726
|
18
|
2,16%
|
13 980
|
12
|
66,67%
|
2,15%
|
30,9359
|
BG
|
53 979
|
47 298
|
6 681
|
3,32%
|
1 623
|
167
|
2,50%
|
3,43%
|
0,7284
|
BR
|
536 740
|
393 644
|
143 096
|
0,71%
|
3 822
|
0
|
0,00%
|
0,97%
|
0,0000
|
CA
|
3 068 237
|
2 693 190
|
375 047
|
0,93%
|
28 436
|
10
|
0,00%
|
1,06%
|
0,0025
|
CH
|
1 064 635
|
1 064 432
|
203
|
1,62%
|
17 283
|
8
|
3,94%
|
1,62%
|
2,4271
|
CN
|
4 532 226
|
4 368 838
|
163 388
|
0,96%
|
43 326
|
1
|
0,00%
|
0,99%
|
0,0006
|
CS
|
166 153
|
165 545
|
608
|
0,02%
|
36
|
0
|
0,00%
|
0,02%
|
0,0000
|
CZ
|
72 516
|
70 453
|
2 063
|
10,02%
|
7 173
|
90
|
4,36%
|
10,18%
|
0,4285
|
DD
|
249 545
|
249 537
|
8
|
0,00%
|
0
|
0
|
0,00%
|
0,00%
|
|
DE
|
7 036 976
|
6 477 194
|
559 782
|
18,59%
|
1 298 948
|
9 462
|
1,69%
|
20,05%
|
0,0843
|
DK
|
448 519
|
393 447
|
55 072
|
3,93%
|
17 536
|
81
|
0,15%
|
4,46%
|
0,0330
|
EP
|
2 795 472
|
1 834 845
|
960 627
|
65,34%
|
1 593 930
|
232 581
|
24,21%
|
86,87%
|
0,2787
|
ES
|
960 530
|
957 819
|
2 711
|
4,86%
|
46 660
|
38
|
1,40%
|
4,87%
|
0,2877
|
FI
|
279 071
|
258 406
|
20 665
|
6,87%
|
19 159
|
4
|
0,02%
|
7,41%
|
0,0026
|
FR
|
3 088 647
|
3 088 557
|
90
|
18,99%
|
586 476
|
66
|
73,33%
|
18,99%
|
3,8620
|
GB
|
3 397 833
|
3 367 739
|
30 094
|
11,14%
|
350 849
|
27 580
|
91,65%
|
10,42%
|
8,7970
|
GR
|
98 917
|
90 240
|
8 677
|
3,46%
|
3 416
|
6
|
0,07%
|
3,79%
|
0,0183
|
HK
|
80 611
|
51 452
|
29 159
|
0,00%
|
0
|
0
|
0,00%
|
0,00%
|
|
HU
|
139 324
|
122 141
|
17 183
|
1,95%
|
2 722
|
1
|
0,01%
|
2,23%
|
0,0026
|
IB
|
86 276
|
86 240
|
36
|
93,56%
|
80 685
|
36
|
100,00%
|
93,56%
|
1,0688
|
IE
|
92 611
|
92 539
|
72
|
2,19%
|
2 027
|
0
|
0,00%
|
2,19%
|
0,0000
|
IL
|
206 937
|
140 258
|
66 679
|
7,48%
|
15 444
|
35
|
0,05%
|
11,01%
|
0,0048
|
IN
|
72 990
|
72 267
|
723
|
7,12%
|
5 200
|
0
|
0,00%
|
7,20%
|
0,0000
|
IT
|
748 377
|
748 298
|
79
|
3,68%
|
27 538
|
0
|
0,00%
|
3,68%
|
0,0000
|
JP
|
17 822 543
|
17 257 179
|
565 364
|
18,04%
|
3 111 474
|
104 145
|
18,42%
|
18,03%
|
1,0217
|
KR
|
2 296 485
|
2 281 075
|
15 410
|
7,28%
|
166 631
|
569
|
3,69%
|
7,30%
|
0,5055
|
LU
|
68 777
|
68 433
|
344
|
0,77%
|
529
|
0
|
0,00%
|
0,77%
|
0,0000
|
MX
|
203 740
|
97 048
|
106 692
|
0,67%
|
1 356
|
0
|
0,00%
|
1,40%
|
0,0000
|
NL
|
615 547
|
614 340
|
1 207
|
6,57%
|
40 409
|
31
|
2,57%
|
6,58%
|
0,3905
|
NO
|
232 604
|
170 575
|
62 029
|
3,56%
|
8 276
|
2
|
0,00%
|
4,85%
|
0,0007
|
NZ
|
122 646
|
82 482
|
40 164
|
3,62%
|
4 444
|
1
|
0,00%
|
5,39%
|
0,0005
|
PL
|
248 692
|
224 078
|
24 614
|
0,57%
|
1 407
|
0
|
0,00%
|
0,63%
|
0,0000
|
PT
|
94 679
|
94 218
|
461
|
0,48%
|
453
|
0
|
0,00%
|
0,48%
|
0,0000
|
RO
|
68 383
|
66 222
|
2 161
|
0,47%
|
324
|
0
|
0,00%
|
0,49%
|
0,0000
|
RU
|
525 910
|
521 178
|
4 732
|
1,72%
|
9 023
|
7
|
0,15%
|
1,73%
|
0,0854
|
SE
|
1 250 035
|
1 246 434
|
3 601
|
3,50%
|
43 763
|
13
|
0,36%
|
3,51%
|
0,1028
|
SG
|
61 904
|
58 124
|
3 780
|
25,61%
|
15 822
|
30
|
0,79%
|
27,22%
|
0,0292
|
SU
|
1 249 768
|
1 248 369
|
1 399
|
0,15%
|
1 937
|
0
|
0,00%
|
0,16%
|
0,0000
|
TR
|
59 895
|
59 626
|
269
|
7,36%
|
4 402
|
5
|
1,86%
|
7,38%
|
0,2518
|
TW
|
636 359
|
634 367
|
1 992
|
0,00%
|
0
|
0
|
0,00%
|
0,00%
|
|
UA
|
51 034
|
46 331
|
4 703
|
1,56%
|
795
|
0
|
0,00%
|
1,72%
|
0,0000
|
US
|
12 720 926
|
12 057 883
|
663 043
|
48,37%
|
5 845 295
|
307 339
|
46,35%
|
48,48%
|
0,9562
|
ZA
|
274 516
|
274 318
|
198
|
0,61%
|
1 686
|
0
|
0,00%
|
0,61%
|
0
|
SOME CONCLUSIONS:
from above data we may say:
1 Coverage of backard citations in national offices is very poor (apart from USPTO) while regional offices (EPO, International Bureau) perform well
2 In most cases backward citations should be enriched with WO equivalent citations (WO/ NON WO ratio < < 1)
thanks to Lorenzo Cassi for rising the issue
No comments:
Post a Comment