Language Model

[Anuj Kumar]

I have a language model, and I want to obtain the list of alternative words
(along with their scores) when I input a phrase.

For instance, if I feed in the word "Good" in a n-gram language model, it
outputs possible next words as "Morning", "Afternoon", "Evening", "Night",
etc. in a "conversations-based" language model.

Would anyone know how to do that?

[Nickolay V. Shmyrev]

The algorithm to select this is trivial, you just check all possible word
combinations and select the best follower

String w1 = "good";
String bestWord = null;
float bestProb = -10000;
for (Word w : words) {
      if (languageModel.prob(w1, w) > bestProb) {
            bestProb = languageModel.prob(w1, w);
            bestWord = w;
     }
}

If you want to do this quickly and repeatedly, you need to precompile this
information into a separate structure.

[Nickolay V. Shmyrev]

And please next time when you post something try to select a more descriptive
subject. Something like "Prediction of most probable following word with a
language model" is way better name of the topic than "Language Model".