Citation Table

Return to Patent Data Specifications.

Table Structure

Table "public.citation"
Column |  Type  | Modifiers 
--------+--------+-----------
patent | bigint | 
cited  | bigint |

Example of table entries:

patent  |  cited  
--------+---------
9226898 | 4935450
9226898 | 5069936
9226898 | 6262019
9226898 | 6264988
9226898 | 6916795
9226898 |        
9226898 |        
9226898 |        
9226898 |        
9226898 |        
9226898 |        
9226899 | 7858080
9226899 |        
9226899 |        
9226899 |

Table Variables

Patent refers to a patent number. Cited refers to a patent number cited by the patent in question. When a patent is issued, the patent can include citations or references to patents in a related field or that may be considered prior art. The table has a composite key that consists of a unique combination of patent number and cited patent number.

Table Purpose

The US patent system allows for inventors, patent attorneys, and USPTO patent examiners to reference other patents that are considered prior art or related inventions. These references are theoretically in place to limit or help define the scope of a patent's claims. Often times, the patent improves upon these referenced patents or may be a new application of previously patented subject matter. Citations may also be unrelated to the patent's subject matter, and therefore, cannot be fully relied upon in studies of innovation and transfer of ideas between inventors. For the Little Guy Academic Paper the citation data may be used as a measure of litigation risk for a patent, as suggested previously by a paper by Lanjouw Schankerman studying patent litigation of small firms. A patent may be more likely to be litigated if it cites many patents or a patent may be at higher risk if it does not cite related subject matter as other inventors may consider this infringement. Our hope is to further investigate the importance and effects of patent citations and define their role in patent litigation and dissemination of knowledge between inventors and IP centered firms. Additionally, the McNair center plans on defining whether or not citations characterize patent quality, whether a patent is useful and novel.

Current Problems

The table has two columns, 'citingpatentnumber' and 'citedpatentnumber'. There are rows with 'citedpatentnumber' greater than 10000000. For instance:

 citingpatentnumber | citedpatentnumber 
--------------------+-------------------
           9226901 |      102005013726
           9226905 |         101332187
           9226905 |        2006528175
           9226905 |        2011513304
           9226905 |     1020090061010
           9226905 |     1020110049808
           9226905 |        2010126349
           9226909 |         101340916
           9226909 |          10128910
           9226914 |         102318827
           9226914 |        2014109862
           9226915 |          10111049
           9226918 |        2008005345
           9226918 |        2008005345
           9226918 |        2008077092
           9226918 |        2008077092
           9226918 |        2008070268
           9226918 |        2008128126
           9226921 |        2008129994
           9226922 |        2012000595
           9226922 |        2012058127
           9226923 |        2010135524
           9226930 |         102002040
           9226930 |          19507522
           9226930 |          50106981
           9226930 |          53082783
           9226930 |          57052334
           9226930 |          10029979
           9226930 |          10045750
           9226931 |        2008203212
           9226931 |        2010227111
           9226931 |        2012068515
           9226931 |        2013106565
           9226932 |         103458880
           9226933 |        2006257751
           9226933 |         101366734
           9226933 |         101396048
           9226933 |         101926831
           9226933 |          20202562
           9226933 |      202005009120
           9226933 |          61063618
           9226937 |          11510473
           9226937 |        2001506579
           9226937 |        2001519791
           9226937 |        2003119127
           9226937 |        2003517831
           9226937 |        2005099761
           9226937 |        2006076681
           9226937 |        2006078941
           9226937 |        2006079021
           9226937 |        2006094209
           9226937 |        2006094210
           9226937 |        2006094233
           9226937 |        2006094235

allpatent_clone=# SELECT  COUNT(*) FROM citations WHERE citedpatentnumber>10000000;
  count  
---------
1411140
(1 row)

 allpatent_clone=# SELECT  COUNT(*) FROM citations WHERE citedpatentnumber IS NULL;
  count   
----------
 23516667
(1 row)

allpatent_clone=# SELECT  COUNT(*) FROM citations;
 count   
----------
 97680838
(1 row)

Possible Solution

Jul 8, 2016: The blank citedpatentnumbers were created due to inconsistency between the original type and the type in the citation table (string to integer). The blank entries mostly correspond to publication number, non-U.S. patent number, and non-standardized patent number. The next step would be to recreate the table accounting for these issues. The U.S. publication number could be matched to the publications numbers in the histpatent table and be replaced by the corresponding patent numbers.

The foreign cited patents will be moved to a separate table.

References

<ref name=IEEE > [1] 'Why do Inventors Reference Papers and Patents in their Patent Applications?', IEEE (2010).