Does Kakao Talk's ommission of 'ending' forms in auto-completion suggestions correspond with the frequency of usage by Korean native speakers?

Question

As a non-native speaker, I'm often confused, when typing a verb, why sentence-ending forms aren't among the suggestions.

For example here, 예쁘다 and 예쁩니다 aren't suggested:

and here, 미안합니다 isn't suggested:

Do those suggestions match how likely those forms would be to be used in actual usage (even considering the 'text-speak' context)?

Answer 1

I want to point out two things.

1. It may reflect low frequency of -다 and ㅂ니다 endings coming right after the verb base form, but it does not necessarily reflect low frequency of "sentence-ending" forms of verbs. For instance, if we look at the suggestions following 예쁘 in OP's first example, we can come up with bazillion ways to "end" the sentence, following any of those suggestions. Some non-exhaustive examples are;

   예쁘

   • 예쁘지(.,?)
     • 예쁘지요(.,?)
     • 예쁘지않(neg)
       • ...
     • ...
   • 예쁜
     • 예쁜데(., ?)
       • 예쁜데요(.,?)
       • ...
     • 예쁜걸(.,?)
     • 예쁜지(.,?)
     • ...
   • 예쁘다는 (. in chat/text context)
     • 예쁘다는데(.,?)
       • 예쁘다는데요(.,?)
       • ...
     • 예쁘다는걸(.,?)
     • 예쁘다는지(.)
     • ...
   • 예쁠
     • 예쁠지(?)
       • 예쁠지도(.)
       • ...
     • 예쁠까(?)
       • 예쁠까요(?)
       • ...
     • 예쁠걸(.,?)
     • ...
   • 예뻐진
     • 예뻐진다(.,?)
     • 예뻐진데(.,?)
     • ...
   • 예뻐지는
     • 예뻐지는데(.,?)
       • 예뻐지는데요(.,?)
       • ...
     • 예뻐지는걸(.,?)
     • 예뻐지는지(.)
     • ...

   (a period in parenthesis means the ending is declarative, while a questions mark means it's interrogative)

   So, yes, sentences end with "a verb stem + 다 or ㅂ니다" combinations, in my opinion as a native speaker, seem to be very rare, not only in chat/text context but also in general, and I believe it's largely because they lack any tense, mood, modal, or any other useful information that more complex combinations of tails can carry.

2. Which leads to the second point that n-gram based word prediction just does not work for morphologically rich languages like the Korean language, given the complexity of the possible combinations that depends on larger context.