Multinomial Naive Bayes and Rapid Automatic Keywords Extraction for Taharah (Purify) Law Chatbot

Rizkhita, Habib Muhtar and Yana Aditia, Gerhana and Dian Sa'adillah, Maylawati and Cepy, Slamet and Cecep Nurul, Alam and Wahyudin, Darmalaksana and Muhammad Ali, Ramdhani (2021) Multinomial Naive Bayes and Rapid Automatic Keywords Extraction for Taharah (Purify) Law Chatbot. In: ICONISTECH 2019, 11-12 July, Bandung.

Text eai.11-7-2019.2298028.pdf Download (535kB)

Abstract

The aim of this study is to utilize the Natural Language Processing (NLP) technology, one of them is in the form of a chatbot. Chatbot has the ability to answer the questions as a conversational search engine. The methods that used on chatbot’s machine are Multinomial Naïve Bayes (MNB) with TF-IDF vectorization to classify the intent, and Rapid Automatic Keywords Extraction (RAKE) to classify the entity. The methods are implemented for thaharah (purify) law as one of Muslim's daily life that cannot be separated from Islamic law. It is important for Muslims to know the thaharah law. The experiments of the methods against chatbot have used a total of 132 data trains and 44 data tests. Results represented by the Confusion Matrix showed the implementation of methods has the overall accuracy 97% with an average precision 90% and recall 97%, which means MNB and RAKE can give the answer well.

Actions (login required)

View Item