The U.S. Patent and Trademark Office granted patent number 9,787,634 to the researchers of the largest domain registrar GoDaddy for “Suggesting domain names based on recognized user patterns”. There are many algorithms for suggesting domain names; however, according to the inventors of this one they mainly rely on “static language dictionaries”. This approach mainly suggests the search of synonyms to key words chosen by the user. However, it neither takes into account the popularity of suggested synonyms as domain names nor the preferences of the user herself/himself.
Suggested dynamic approach should solve these tasks, reports Domain Name Wire. For instance, the algorithms they created will automatically filter out the words that are unpopular when it comes to domain names. The program will also be able to analyze previous requests of each user. For example, if a registrant already tried to find a domain name and subsequently sorted out the options (by adding letters, numbers of other words to the key word), then all of these searchers will be taken into account. As a result, the user should get a list of names not only potentially popular, but also that corresponds to his or her preferences. Algorithms of machine learning will constantly improve data processing and improve the quality of the results.
