A CONVERSATIONAL SYSTEM (Chatbot) FOR HANDLING ADMINISTRATION-RELATED STUDENTS’ QUERIES

ABSTRACT

The administrative challenge of efficiently answering the questions of several thousands of students is one that tertiary institutions are looking for innovative ways to solve. There seems to be no shortage of enquiries and queries, leaving the student affairs administrative unit overwhelmed with repetitive work. This Research study details the creation of a chatbot called CovenantBot. This text-based conversational system serves as a first-line approach to provide students with responses to their school-related queries. Matrix Laboratory (MATLAB) was used to create the chatbot. The chatbot was implemented using the rule-based approach, where pattern recognition was used to trigger an entire set of responses from the chatbot. The system's knowledge base was created in Microsoft Excel and parsed into its essential parts, and the keywords were extracted and used to train the chatbot. The word frequency evaluation of the proposed CovenantBot showed that questions related to 'student' had the highest frequency. The accuracy, precision, and recall of the proposed CovenantBot were 0.66, 0.74, and 0.67. The resulting reactive chatbot was able to provide satisfactory answers to administration-related queries from students.

Keywords: Chatbot, conversation, rule-based, MATLAB, Natural language processing, Pattern recognition, Student, Administration, Queries, Database, Tertiary institution.