US Patent Classification System
|This Glossary entry exists for the community to share information related to common terms used in prior art searching. Registered users can add, edit, or delete material on this page. Users should keep in mind that the information on this page is the result of community collaboration and, as such, is vetted by the community at large, not individual experts or fact-checkers. All information contributed to this page is public information - do not post confidential information. For more information about creating and editing Glossary articles, please see our Help pages. If you found this page through a web search, we invite you to visit our Main Page to see what Intellogist is all about.|
Although it is only applied to United States patent documents, the US Patent Classification system is one of the most important national patent classification systems. It is heavily used today, because the economic importance of the US patent system makes US patents a vital source for many prior art searches around the world.
There is often no exact relationship between International Patent Classification (IPC) and US marks since, unlike many other national/regional use classification systems such as DEKLA (used by the German Patent Office) and ECLA (used by the European Patent Office), the current US classification system has no relationship to the International Patent Classification (IPC) system. The US system was developed in 1836, whereas the IPC system was first introduced over 100 years later, in 1968.
Advantages to Using the US Patent Classification System
There are several advantages to using the US patent classification system when searching US documents. First, it is perceived that US patent examiners classify patents with US marks more accurately than they do with IPC marks. This perception is widely considered to be valid because it is known that US examiners classify using the US system first, and then often use a US to IPC concordance tool to automatically generate IPC codes; as with any machine-generated data, and because there is not an exact 1:1 relationship between US and IPC classifications, the IPC concordance tool can sometimes produce IPC marks that are not optimal. Although US examiners have the IPC class definitions at their disposal to check the accuracy of the concordance, it is presumed that not all examiners take the time to do so, since the USPTO is widely known to emphasize the need for high case output, and many examiners often struggle to meet production needs.
In addition, it was the perception in the field at one time that the US classification system was revised much more frequently than the IPC system, and consequently that it was quicker when adapting to changing and emerging technologies. Since the introduction of the IPC-Reform system in 2006, however, this advantage may no longer be reality, as the IPC-advanced designations are supposedly revised and updated every 3 months. Representatives of the United States Patent and Trademark Office's (USPTO) Office of Patent Classification were not able to give statistics on how often the US classes are currently revised, but it is clear that the revisions take place on an as-needed basis, with examiners or teams of examiners usually making the recommendations for how the system should be revised. Once a revision is proposed and approved by the Office of Patent Classification, the entire backfile of US patents is re-classified in accordance with the revisions. The re-classification work is done manually by evaluating the content of the patent claims; it is performed by a private company contracted by the USPTO.
Today, thanks to revision and re-classification efforts, the current US patent classification system can be used to accurately search US granted patents back to the 1790s. Although the classification system was first introduced in 1836, it appears that a backfile re-classification project has extended it back to the 1790s, as sources at the British Library website and the USPTO search site both state that the system extends to those early dates. Due to the length of use of the system, however, it does contain some dead weight sections, and the numerical order of the classification system has become somewhat jumbled, with certain sections having subclasses extending out into two and even three decimal points, while other sections have wide numerical gaps. While the length of the system means that it can be used quite far back historically, the numerical confusion that has been introduced over the years is definitely a drawback to the system. More on the potential numerical pitfalls of the US class system is discussed below.
Anatomy of a US Classification Symbol
Classification is a hybrid of “functional” classes (focused on an aspect of an invention) and “application” classes (focused on an industry) US classifications have two substituent parts: they are broken into Class and Subclass, where the class is a number up to 3 digits, and the subclass is a number up to 6 digits where the last three digits are decimal places. Subclasses are usually 1 to 3 digit numbers, and subclass numbers with more than one decimal place are rare. Usually, the class and subclass numbers are separated by a slash, although some electronic databases will require users to remove all punctuation (including the decimal point in the subclass number, if it is present).
The number format follows:
- Class number/Subclass number (Subclasses are usually 3 digits, but can be appended through the use of decimal points)
Because electronic databases usually require the user to leave out the slash and decimal place, the classification must be “filled out” out to the appropriate digits by using leading zeros (such as 007 for class 7). Because of the way the data is formatted by the USPTO, sometimes in-system data points will also have trailing zeros (for example, placing 000 at the end of a 3 digit subclass to represent the missing “sss” decimal spots), but most databases will clean these up before they are entered into the database. It is always important to check system recommendations before formatting a US class search query.
One of the things that makes the US classification system confusing is that subclasses are organized into headings and sub-headings within themselves. There are major subclasses (or mainline headings) along with up to 6 levels of sub-headings under these mainline headings. Unfortunately due to the long development time and many revisions of the US class system, subclass numbering is not intuitive – there is no way to tell the hierarchy or level of indentation of a given subclass by looking at its number. For example, review the following series of subclass numbers under class 62, refrigeration, and try to determine which are major headings and which are subheadings within the subclasses: 45.1 46.1 46.2 46.3 47.1
The correct heading level is shown via the dots prior to the subclass number:
Without an in-depth knowledge of class 62, no random observer could have guessed that 46.1 and 47.1 were indented one place beyond the major heading, 45.1. The fact is that these numbers are very similar, and that there is no way to tell whether they are all equal, or which are likely to be indented more than the others. Also of note is that design patents have their own “D” classes.
Searching Using US Classification
There are two tools available on the USPTO website to help users identify US classification areas of search:
- The Manual of Classification: this interactive tool shows classes in their hierarchical organization, with short definitions as they relate to sub and super-level classes
- The Index to the US Patent Classification System: this tool presents a list of keywords in alphabetical order, and lists the classes related to the keyword concepts.
A third, very effective method to identify relevant classes is to do a title/abstract keyword search for the concept the searcher is looking for, browse a few results, and then study the classes printed on relevant documents for a hint.
Again, when searching US classifications in an online database, it is important to be aware of the query requirements. Some databases have autoposted US data at the class level, meaning that entering just the class number in the US Classification search field will search for all patents in that class, regardless of subclass designation. However, this technique will not work for subclasses: entering 62/45.1 into most search databases will only find documents with the 45.1 term, but will NOT find sub-level designations in the hierarchy such as 46.1-46.3, 47.1, etc. For this reason it is always important to investigate a class/subclass found from a quick patent search, to determine if there are other sub- or super-level classifications that are also related to the relevant search concepts, since these should be searched as well.
To circumvent the problem described above, there are some tools that can be used to generate automatic lists of subordinate classes. Because classification hierarchy is not discernable from subclass numbering, these tools always have to rely on the USPTO’s published Manual of Classification to determine classification hierarchy (such as ClassClarify). They are performing work that could be done for free, but they may be of use to searchers without much time to spare. The bottom line is that some background work is usually needed in order to search a classification system effectively.
Automatic Patent Classification
In addition to looking up the right patent classification using the USPTO-provided tools (described above), there are third-party solutions, such as PatentClassifier.com that suggests a US classification class and subclass based on the text a user provides. According to the PatentClassifier.com website, the accuracy of automatic classification was determined to be in the 80-90% range, depending on the subject matter.
US to IPC Concordance
An automatic concordance tool to convert US classes to International Patent Classification (IPC) classes is also available on the USPTO website. This can be useful for hints but should not to be relied upon solely, since the systems are so different that it is impossible to get a 1 to 1 concordance correct. Harmonization projects between the USPTO, the European Patent Office (EPO), and the Japanese Patent Office (JPO) are sometimes conducted to re-arrange US classes to correspond to the IPC system in a 1 to 1 relationship. However, this has only been performed for certain portions of the US system, such as parts of class 588.
Additional Types of Art Collections
The US Patent Classification System has several special case subclassifications designed to organize patents into categories that relate to inter-class, proto-class, and foreign-class designations. The following are additional types of art collections as specified in the USPTO Handbook of Classification.
Digests are collections of related patent documents that fall into a sub-category that may span across several subclassificatons. The classification 237 DIG1, for example, is a collection of patent documents that fall under the 237 class (Heating Systems) and are all related to Inert Gasses. Digest classifications are not allowed to be the primary classification of a patent document; assigning a patent document to a digest as a cross-reference is discretionary. Digests are officially undefined.
Cross-Reference Art Collection
A Cross-Reference Art Collection is similar to a digest with the exceptions that Cross-Reference Art Collections are officially defined and generally take up the 900-999 subclass range when available. Cross-Reference Art Collections are not allowed to be the primary classification of a patent document; assigning a patent document to such a collection as a cross-reference is discretionary.
Alpha Subclasses Alpha Subclasses are proto-classes developed from within an existing subclass as a way to facilitate classification searching. Patent documents are selected and removed from the original subclass to populate the newly created alpha subclass. This leaves the original subclass without many patent documents, and therefore an "R" suffix is added to make note of this. Multiple alpha subclasses can be created from one original subclass.
There are no official definitions for alpha subclasses, but they receive definitions if they are optionally later converted to regular subclasses. Alpha subclasses may be used for all types of classifications including primary classifications on patent documents.
E-subclasses were created in September 2002 to "allow integration of European Classifications (ECLA) into USPC classes for enhanced search capability."
E-subclasses most often correspond to an ECLA subclass on a one to one basis and contain US and foreign patent documentations. E-subclasses that are suffixed with EPO or JPO are automatically and regularly updated with patent documents from the respective cited source. As part of the E-subclass initiative, anyone with questions can email E-Subclasses@uspto.gov.
FOR subclasses are collections of the foreign patent documents that were reclassified by the USPTO prior to 1995, a practice which now only exists on a limited basis. The FOR subclasses contain only foreign patent documents and non-patent literature. When a US subclass is abolished, the foreign patent documents are not reclassified into new US subclasses. Rather, the foreign patent documents are allocated to FOR subclasses. The parenthetical number trailing each FOR subclass is the corresponding abolished US Patent Classification code.
- ↑ "Important Notices concerning the Patent Full-Text and Full-Page Image Databases." USPTO, http://www.uspto.gov/patft/help/notices.htm. Accessed on November 18, 2008.
- ↑ "Accuracy of the classifier." http://patentclassifier.com/learn/accuracy.html. Accessed on December 29, 2014.
- ↑ "Handbook of Classification." March 2005. USPTO website, http://www.uspto.gov/web/offices/opc/documents/handbook.pdf. Accessed on November 28, 2008.
- ↑ "E-Subclasses (ECLA-Derived Subclasses)." September 23, 2002. USPTO website, http://www.uspto.gov/go/classification/bulletins/bulletin_g.htm. Accessed on November 28, 2008.