How to start a patent search?
Step 1: List down the keywords
The first step for patent search is in-depth understanding of the proposed invention/ idea and identifying the most important features of the particular invention. After the features have been identified, the searcher has to list down the most important keywords from the features. Keywords are the search terms used for performing a patent search. Selecting correct and relevant search terms is very important for removing irrelevant patents from the result. Also the search terms should not be very common and should be according to the key features of the proposed invention. Identification of synonyms of the keywords should be done for good search results.
Step 2: Finalizing the database(s)
After listing of the keywords for the proposed invention, the next step is the finalization of database(s). There are many databases available online for performing patent search. One or more than one database may be chosen for the process. The databases can be the repository of the respective patent office’s like USPTO for USA patents/patent applications, Espacenet for European patents, enterprise databases such as Google Patents, Free patents online, Thomson Innovation, patbase, and the like.
Step 3: Perform keyword based search by preparing search strings
A key string is a combination of keywords related to the invention for performing keyword based search. A key string is entered into a patent database. Key strings can be formed to map key features in various parts of a patent like Abstract, Title, Claims, Description or even combination of all four. Every database has a particular syntax for writing a search string or key string. For example PATSEER database allows one to combine keywords along with their synonyms by just using an OR operator and also allows the searcher to join the combinations of the words using AND operator.
For example- a key string can be ((mobile OR phone OR “communication device” OR “calling device”) AND (design* OR figure OR pattern OR blueprint))
The key strings defined for the invention may be of two categories i.e. broad key strings and narrow key strings. Broad search strings are those strings which provide a wide range of results to the searcher whereas Narrow search strings provide limited but focused results to the searcher. Generally, there is greater number of keyword combinations in a narrow search string for providing limited but accurate results whereas a broad search string utilizes less number of keyword combinations for a larger group of results. The narrow search string is employed by the searcher if he isn’t able to obtain relevant results with a broad search string.
Broad search string
(mobile OR phone OR “communication device” OR “calling device” ) AND (design* OR figure OR pattern OR blueprint)
Narrow search string
((mobile OR “communication device” OR “portable device” OR phone OR “calling device”) AND (“communication system” OR network OR PSTN OR PLMN OR GSM OR “cellular system” OR “public switch telephone network” OR “public land mobile network” OR “global system for mobile communication”) AND (“subscri*” OR “register*” OR agreement OR “enroll*” OR “membership*”) AND (“inform*” OR content OR data OR “detail*”) AND (inactive OR off OR sleep OR idle) AND (mode OR “operate*” OR “stat*” or “function*”) AND (“select*” OR choose OR “prefer*”) AND (“service area” OR location))
Step 4: Extract patents and screening all the patents for identifying relevant patents
The next step for patent search is to study all the extracted patents and filter the relevant patents. After forming and entering the particular search string, the searcher has to extract search result containing patents and study them. Each extracted patent is examined carefully and the patents which are found to resemble the key features of the proposed invention are filtered out and are identified as relevant patents. This process might take a long time or less time depending on the quality of the results and understanding of the proposed idea. Some patents may found to have some features of the invention whereas some patents may have all the features related to the invention. Therefore, these relevant patents can be divided into various categories i.e. highly relevant, Relevant and Related. The highly relevant patents will include a list of those patents which have all the key features of the invention.
Identify key inventors and assignee
For all the extracted relevant patents, inventor based search or assignee based search should be performed. The inventor based search is the search for the patents according to the inventor or the assignee to which the patent/patent application has been issued/filed. The inventor or the assignee can be either one or it can be a group of people or a company. For example, a patent for software can be searched under the patents issued to Microsoft or any other software organization.
The assignee based search can be performed independently or in conjunction with keywords. Performing assignee based search independently requires directly searching for the patents according to the assignee name to which the patent has been issued. In addition, if an assignee or the inventor has several patents issued to his name that belongs to different technological domains, then the searching performed is called assignee based searching in conjunction with keywords. For example, if an inventor has issued patents in mechanical field, biotechnology field, electrical and electronics field, then the searching is performed according to the keywords related to these fields.
Perform spider search and citation on patents set 1 – patent set 2
Citation in terms of the patent means a reference to prior arts. Citations can be previous patents or they can also be journals, articles, published papers and literature. Citation search is done when a searcher is not satisfied by the results or the number of hits or he wants to make sure that he hasn’t missed out on patents which might be relevant. Therefore, to increase the scope of the patent search, citation searching of the patents which are found relevant based on the keyword-based search or another type of search is done. Citation searching will help obtain results which would not have been produced by previous search techniques. Citations can be backward citations or forward citations. A backward citation is a document published earlier than the current document and the current document is the forward citation of the earlier published document. Citation search is used to enhance the search strategy. For performing the citation search, first, extract the citations from all the relevant patents found during the initial search. Thereafter, analyze the cited documents to find out more relevant references. If the searcher found more relevant patents from the analyzed citations, then it’s good to perform citation search on the newly found relevant patents to further explore more relevant documents.
Perform date based search and other combination of searches
Date based search in a patent search is a process of extracting relevant patents by making use of the various dates mentioned in the patent. Each patent has various dates associated with it like filing date, publication date, and priority date. A searcher can perform a patent search based on these dates for finding relevant patents. Patent searching databases give the option of performing the search by mentioning a particular date or a range of time within which you want to find out the relevant patents. For example, if one has to extract the relevant patents before or after a particular date, he can easily mention the date in the date column. This feature of the database(s) helps to narrow down the search results. The searcher can extract patents before or after the priority date, filing date, publication date or issue date.
In addition, the searcher can also perform various combinations of searches for narrowing down the search results. Various combinations can in the search string or key string can be made with the priority date or date along with the assignee’s name, filing date and inventor name. Therefore, there can be many permutations and combinations for performing a patent search. Some patent databases also provide the option of searching through key strings in “title + abstract” or “title + abstract + claims” or “title + abstract + claims + description” which helps in finding more accurate results. Moreover, these combinations can be grouped with other types of searches to form another combination which can provide specific results.
Non-Patent Literature search
Searching can also be done for non-patent literature. The non-patent literatures include IEEE papers, ACM, ACM Digital Library, ACS Publications, AIP Advances, Agricola, ArXiv, BioAlma, BioMed Central, BioMedSearch, Books24x7, CHEMCATS, CHEMLIST, Cases Database, Chemistry Central, Chimica, CiteSeer, CiteSeerX, Collaborative Drug Discovery, Computer Science Digital Library, EnCompassLit, Environmental Sustainability, Espresso, Europe PubMed Central, Gallica Digital Library, Geobase, Global IP Law Service, Google Scholar, Hakia PUBMED, HealthStats, LexMachina, LexisNexis Dossier, LexisONE, Libdex, Library of Congress Website, Libweb, LitAlert, M-CAM, MathSci, MedTRACK, Merck Index, MetaPress, Microsoft Academic Search, NCBI BLAST, NCBI Entrez, NTIS, National Library of Medicine Catalog, Nature, NetLibrary, NextBio, IBM Technical Disclosure Bulletins, IEEE Xplore, IHS Engineering360, IP Litigation Clearinghouse, IPQuester, ISI Web of Knowledge, InfoSci-Books, Infotrieve Online, Ingentaconnect, Inspec Direct, Intute, PASCAL, PharmaLiveSearch, PubChem, PubMed Central, PubMed Central Canada, PubMed NCBI, OCLC Libraries Worldcat, OnePetro, Online Mendelian Inheritance in Man, OpenThesis, ScienceDirect, Scopus, Social Science Research Network (SSRN), Software Patent Institute (SPI), Springer Open, Springer Protocols, SpringerLink, SpringerReference, etc.
Further, similar to searching on the patent databases, there are different ways to search the non-patent literature. These include keyword-based search and author based search.
Most of the searcher begins their searching by using a few keywords from the proposed invention. The searcher goes to any of the free non-patent databases and types the keywords from the invention. So, the searcher may not go through the entire list of the non-patent literature, but only a list of non-patent literature related to his area of interest. The keywords should be such that it describes:
1) Purpose of the invention
2) Proximate or essential function of the invention
4) End result
One issue with the keyword-based searches is that the inventor’s semantics may not match with the semantics or terminologies of the others. The author is allowed to create his own terminology to describe the invention. For example, a mouse trap can also be used as a ‘rodent extermination device’ by some authors. Another issue is that most of the US publications are from other nations. The foreign entity publications may contain misspelled words e.g. colour in place of color. However, these errors can be rectified but the rectification process is time-consuming. This is the yet another issue with the keyword-based searches.
Author based search
One can also start the patent search by using the author’s name. It is almost similar to inventor based search in case of patents. One can directly search on the web or use the database(s) provided for searching.