Chatbot for Virtual Travel Assistant with Random Forest and Rapid Automatic Keyword Extraction

Abstract

The purpose of this research is to create a web- based Chatbot application by implementing the Random Forest (RF) algorithm and Rapid Automatic Keyword Extraction (RAKE). The chatbot application will be used to become a virtual tour guide regarding tourist objects in Bandung, Indonesia. This is because the need for information is very important for most people to help and accelerate the work to be done. The data used in this research is a set of data in the form of questions as many as 192 data. The dataset is divided into training and testing data of 150 and 42 data. RF algorithm will work as an intent classifier, and RAKE will work as an entity recognizer to determine the answer to the user. The classification model is evaluated using training and testing data scenarios after the pre-processing process, such as case folding, removing regular expression, word tokenizing, and stop-words removal. The evaluation using the confusion matrix shows that an average accuracy value of the chatbot application using RF and RAKE is 98%, the precision value is 96%, and a recall value is 96%.