The chatbot application can be used to help provide infomration needs just by sending a texc message to the bot. In this study, chatbot is used as an information service for informatics engineering majors at PLN Technical College
Jakarta. Named Entity Recognition (NER) is used to identify keywords from human language. The keyword pattern is matched to a pattern in AIML (Artificial Intelligence Markup Language) to provide relevant answers. The named entity
recognition of words in this study using Multinomial Naïve Bayes algorithm. Broadly speaking, there are two processes performed, namely the process of NER learning and chatbot process. Before performing the calculations, the text first passes the tokenizing process, part of speech tagging, BILOU tagging and feature extraction. Based on the test results, this chatbot app can recognize the named entity of a word with an average precision value of 88% and an average recall value of 85% and can provide a precise answer with an accuracy of up to 82% based on a pattern already found.Based test results, this chatbot app can recognize the named entity of a word with an average precision value of 88% and an average recall rate of 85% and can provide a precise answer with an accuracy of up to 82% based on a pattern already found.

Aplikasi chatbot dapat digunakan untuk membantu memberikan kebutuhan informasi hanya dengan cara mengirimkan pesan teks ke bot. Pada penelitian ini, chatbot digunakan sebagai layanan infromasi bagi jurusan Teknik informatika di Sekolah Tinggi Teknik PLN Jakarta. Named Entity Recognition (NER) digunakan untuk mengenali kata kunci dari bahasa manusia. Pola kata kunci tersebut dicocokkan dengan pola di AIML (Artificial Intelligence Markup Language) untuk memberikan jawaban yang relevan. Pengenalan named entity pada kata dalam penelitian ini menggunakan metode Multinomial Naïve Bayes. Secara garis besar, ada dua proses yang dilakukan, yaitu proses NER learning dan proses chatbot. Sebelum melakukan perhitungan, teks terlebih dahulu melewati proses tokenizing, part of speech tagging, BILOU tagging dan feature extraction. Berdasarkan hasil pengujian, aplikasi chatbot ini dapat mengenali named entity dari suatu kata dengan nilai rata-rata precision mencapai 88% serta rata-rata nilai recall mencapai 85% dan bisa memberikan jawaban yang tepat dengan akurasi hingga 82% berdasarkan pola yang sudah ditemukan.