Which mode/method to use that is agnostic to word orders?

[skwskwskwskw]

Hi,

Would like to understand which matching algo/model is agnostic to word orders? I realised for instance Levenshtein Distance might be affected by word orders.

Thanks

[MaartenGr]

You can use TF-IDF for that since it typically only considers n-grams on a token level. Due to its bag-of-words like approach, it does not take the order of n-grams into account and therefore also not the order of words.