Does Azure Search use synonyms

Can Azure Search be based on synonyms?

eg. If I search for "shoes", the system should get results for shoes synonyms such as shoes, sandals etc

I've searched a lot and couldn't find the correct documentation about this. The closest I could find is a post that says it is not supported yet, but this is an old post - https://social.msdn.microsoft.com/Forums/en-US/9d0873cf-14aa-4d3e -b33f-a047963e8671 / internationalized-search-finding-both-color-and-color? forum = azuresearch

+3


source to share


1 answer


Unfortunately, we have not yet added support for auto synonym to Azure Search. Of course, as you can imagine, this is a pretty important feature that we hear quite a bit, so we have plans in this area. In the meantime, let me give some pointers:

1) We support creation, and in fact the Microsoft Natural Language Processor (NLP) that is part of our preview api 02-28-2015-Preview is really good at lemitization which provides a really deep understanding of 50 languages ​​to allow you to get different forms of words. In English, a good example would be where we can find mice when looking for a mouse. It also does a great job with tenses (masculine / feminine word forms). I understand that this is not what you are talking about in your question, but there might be something you can use. More on this here: https://azure.microsoft.com/en-us/documentation/articles/search-api-2015-02-28-preview/#LanguageSupport

2) If you know the synonyms you want to use, there are several methods you can use to get synonym support in Azure Search, including:



  • Use our collection data type where you view the content of your documents and where there is a possible match for a synonym word, add that synonym word to the collection field for that document. So if someone does a word search for synonyms it will look like a match
  • Create a separate index that includes only these synonym words. When you do a search (or search for offers), it goes against that "synonym" index to get all possible matches, and then makes a matching query against the main index.

I hope this helps

+3


source







All Articles