Dealing with lots of non-dictionary words

Hi

We're just starting to implement ES, moving over from Sphinx

One of the big issues we currently face, is dealing with misspellings. However, in our realm, lots of our documents contain branded words which are not in the dictionary

For example, names of artists, brands, etc.

It's almost a reverse spell check "corruption" => Did you mean "Korruption?" (where Korruption is indexed highly in our documents)

Because some of these are misspelt by the users when searching,we want to suggest the correct spelling, but this must be done according to indexed documents, rather than dictionary terms.

I've been looking at https://www.elastic.co/guide/en/elasticsearch/reference/current/search-suggesters-phrase.html and hoping this will work, perhaps setting ** real_word_error_likelihood** lower?

Any other suggestions on the best implementation would be very welcome

Thanks