Named Entity Recognition on Indonesian Tweets using Hidden Markov Model

Tweets are short messages that can load up to 140 characters, posted by Twitter users. To meet this limit, users usually use abbreviation to express their thoughts, so that produce unstructured and inappropriate sentences grammar. As a result, it is difficult to identify which entity are needed, such as people's names, locations, and organizations. On this research, developed an entity named recognition system on Indonesian-language tweets using Hidden Markov Model with the addition of the POS Tag feature. Using metrics evaluation, the biggest F₁ score is 64.06%. Our source code and data are available at: https://github.com/dirazarine/Named-Entity-Recognition-on-Indonesian-Tweets-using-Hidden-Markov-Model.