Descriptions and definitions (WIPO)

Data set description

The data source is the European Patent Office’s PATSTAT database for all citation information. Bibliographic information for international applications is taken mainly from the PATSTAT database, supplemented by information from WIPO internal databases where information could be provided which was not available from PATSTAT.
The data provided is based on published PCT searches.
Statistics are presented by search date, meaning the date on which an international search report was transmitted to the International Bureau (since this information is available more consistently than the actual date of search).
The date ranges for statistics take into account data availability. This is constrained by procedural latency such as time to publication, as well as cut-off dates for database extracts.
No filing date constraint is applied.

Applications with no citation recorded are removed, as no meaningful international search was carried out for these applications.
A small number of patent citations are without category codes.
In case of citn_origin = 5 (documents cited during international search), those citations are considered; otherwise, citations with citn_origin = 0 (documents cited during search) are considered. Citations with other citn_origin codes are removed.
NPL citations with no category assigned and with ID >= 900000000 are removed, as they don’t seem to be in the original search reports.
All citation category codes recorded in the database for the valid citations are considered.
Citation language codes for national patent documents are those recorded in the Patstat database, citation language codes for PCT documents are assigned using WIPO’s PCT database as they are more precise. The language codes are further cleaned up according to information of the authorities who publish those documents.
No attempt has been made to determine the language of publication of non patent literature documents.

Technology sector and field are derived from the IPC classes assigned in the international phase search report or publication.
The grouping into technology sector and field is based on a concordance provided by WIPO. (http://www.wipo.int/ipstats/en/statistics/patents/pdf/wipo_ipc_technology.pdf).
This technology breakdown includes 8 technology sectors (Electrical engineering, Energy technology, Instruments, Mechanical engineering, Micro-structural and nano-technology, Other fields and Semiconductors), which are further broken down into 35 technology fields.
Multiple IPC classes are often assigned to applications. For the present statistics, fractional counting method is applied, that is, an international application and all citations in its search report are evenly distributed to multiple technological fields when multiple fields are associated with it.
IPC class information is not available for approximately 1% of applications.

In general this is the State in which the first-named applicant is resident (overall, this gives a more useful indication of origin of the application than the receiving Office because the International Bureau and regional Offices work for many States, whereas some States do not themselves operate a receiving Office).
“Unknown” code is used for a small percentage of applications.

XY rate refers to share of search reports where at least one citation is in the category of X or Y.
In addition the use of an E citation is counted as XY if it can be assumed that the E citation is prejudicial to novelty. This is the case unless the E category is assigned in combination with A.

PATSTAT does not contain all citation categories for each citation. The database contains one citation category per group of categories for each citation. The category groups are defined as follows:

Only one category from the same group is selected. The category selected is determined from the order in the table above. In this way a citation will be categorized as X if the citation categories in the search report are XY for this citation. Priority for selecting the letter is according to the ranking of categories left to right within the groups above, rather than the order of their appearance within the citation in the international search report (that is, X will be shown even if the search report lists Y category claims first).
Citations from some search reports are recorded in PATSTAT with an unofficial category “I” indicating a document cited for inventive step on its own, without reference to a second document. These citations are treated as the official category “X”, of which “I” is a subset.
A maximum of 3 categories is recorded.
Citation Category Examples:

Search report citation	Citation categories present in PATSTAT
X, Y, A, P	X, P
Y, A, P, E	Y, P
Y, X, O, T	X, O
X, P, D, O	X, P, D

This means for example in row 1 above the Y nor the A is not stored in PATSTAT.
In practice it is therefore possible to determine whether a search has at least one X or Y citation. It is also possible to correctly count the number of X citations.
In approximately 20% of cases it is not possible to correctly count the number of Y categories used, although it is possible to count the use of Y without an X.
EPO data has been refined with an additional internal data source.

A-only rate refers to the share of search reports where no citation is in the category of X, Y or E.

Y no X rate refers to the share of search reports where at least one citation is in the category of Y and there is no X citation.

The date when the search report is transmitted to WIPO (the actual date of search is not available in all cases).

Citations in PATSTAT are categorised into patent literature and non-patent literature.
A citation is considered patent literature if it relates to patent abstracts provided by various providers.
Less information is available for NPL citations. For example, the language of a NPL citation is not available.

This is used for counting patent citations that are not in an official language of the respective ISA:

ISA	Official language
AT	German
AU	English
BR	Portuguese
CA	English, French
CL	Spanish
CN	Chinese
EG	Arabic, English
EP	English, French, German
ES	Spanish
FI	Finnish
IL	English
IN	English
JP	Japanese
KR	Korean
PH	English
RU	Russian
SE	Swedish
SG	Chinese, English
TR	English, Turkish
UA	English, French, German, Russian, Ukrainian
US	English
XN	Danish, Icelandic, Norwegian
XV	Czech, English, Hungarian, Polish, Slovak

The statistics are based on the actual official languages of the Office, but can easily be redefined to reflect any set of core languages which an Offices considers to be useful in assessing how effective its processes may be at discovering prior art beyond those languages.

This is the patent organization who published a citation document.
It is normally a national patent office, a regional office such as the EPO, or WIPO.

Generally processing authority is assigned from the publication authority of the citation.
For WO publications, the international search authority is chosen to indicate which office processed the cited patent publication. This gives an indication of the nature of the publication which be more useful for some purposes than simply the number of WO citations, which may be in any of 10 languages.