EPO - Discussion forums

Hello,

IF only 50% of the applicants were marked for their country origin and no applicants from Japan included, Does it still have statistical meaning of this index?

I want to draw a grographical map representing the applicant's origin in a global scale. Is ther anyone could tell me it is possible or not? Are all the applicant without specified origin coming from Japan?

thanks!

By origin do you mean nationality or address? If address, then by looking at table `tls206_person` it is obvious that there are many missing country codes. It is also obvious that these aren't just people from (or in) Japan.

If I'm not mistaken there's a post dealing with this issue on http://rawpatentdata.blogspot.com/
http://rawpatentdata.blogspot.com/2010/ ... ntors.html

Now I understand that searching for patents based on applicant and inventor country of origin is difficult because information is not always available due to fields being left blank. I am also aware that not all intellectual property offices provide the same quality and completeness of data.

Seeing as this post is several years old, I was wondering if progress has been made in this area. Is there an approach to solving this issue that is recognized as more appropriate than other and are there any studies that have used this approach?

Any help is much appreciated.
Thank you

There are ways to find the missing data.
For a recent and very good example, see the "worldwide count of priority patents" paper by Rassenfosse et al.

I have come accross the same Problem yesterday, when trying to find out the fractional value of "Chinese applicants" - (a vague concept as was made clear).

First, I opted for tls206_person.person_ctry_code='CN' only - an ideal world assumption.

Then

(tls206_person.person_ctry_code='CN'
OR
tls206_person.person_name ~* 'CN|China'
OR
tls206_person.person_address ~* 'CN|China'
)

and extended the select to INPADOC families persons.


But still, the number I get is much smaller than that published by SIPO.

The full questions read: In which IPC do Chinese applicants file in China, (and the non-Chinese complement)?

No Im very insecure to interpret the result, actually doubting the integrity of my patstat implementation (Postgres)
If may be someone could vaguely check the query? - I am not so fluent/frequent in SQL.


Select

substring(tls209_appln_ipc.ipc_class_symbol from 0 for 4) ,
count( substring(tls209_appln_ipc.ipc_class_symbol from 0 for 4))

from

tls201_appln, tls209_appln_ipc, tls207_pers_appln, tls206_person, tls219_inpadoc_fam f1, tls219_inpadoc_fam f2


where

tls201_appln.appln_auth ='CN'
AND
tls201_appln.appln_id= tls209_appln_ipc.appln_id
AND
f1.appln_id=tls201_appln.appln_id
AND
f1.inpadoc_family_id=f2.inpadoc_family_id
AND
f2.appln_id=tls207_pers_appln.appln_id
AND
tls207_pers_appln.person_id=tls206_person.person_id
AND
tls207_pers_appln.applt_seq_nr>0
AND
(tls206_person.person_ctry_code='CN'
OR
tls206_person.person_name ~* 'CN|China'
OR
tls206_person.person_address ~* 'CN|China'
)
AND
extract(year from tls201_appln.appln_filing_date)>'2010'



GROUP BY

substring(tls209_appln_ipc.ipc_class_symbol from 0 for 4)

ORDER BY
substring(tls209_appln_ipc.ipc_class_symbol from 0 for 4) ASC


;

Have a wonderfull day,
Felix

Try this:
SELECT * FROM `tls201_appln`
INNER JOIN `tls207_pers_appln` ON `tls201_appln`.`appln_id` = `tls207_pers_appln`.`appln_id`
INNER JOIN `tls206_person` ON `tls207_pers_appln`.`person_id` = `tls207_pers_appln`.`person_id`
WHERE `appln_auth`="CN" AND `tls207_pers_appln`.`applt_seq_nr` > 0 AND `tls206_person`.`person_ctry_code` = "";

The idea behind it is as follows: if someone is applying for a patent in China, and the country of the applicant is not given it is probably because the applicant is Chinese. Why add the country for a local kind of thing.

Besides the paper by Rassenfosse there is also a project called APE-INV, though it focuses on inventors instead of applicants. And Europe instead of China. But who knows.