LINE CHATBOT FOR COSMETIC ADVERTISING INFORMATION USING DIALOGFLOW AND GOOGLE SHEETS

Abstract

After the COVID-19 outbreak, economic growth, especially in Thailand’s cosmetics and beauty market, has led to rapid market expansion with significant increases in cosmetic registrations and requests for consultation on advertising cosmetic products. Currently, the Cosmetic and Hazardous Substances Advertising Group, Food and Drug Administration services has insufficient staff to provide information and to reply to repetitive inquiries. The objectives of this research are: 1) developing a LINE chatbot to provide information on cosmetic advertising for entrepreneurs; 2) evaluating its accuracy; and 3) assessing LINE chatbot user satisfaction. This study follows a research and development methodology, comprising the process of designing and developing a LINE chatbot using Dialogflow, working with the Line Messaging API and Google sheets. Results showed that the developed LINE chatbot is effective, as assessed by Confusion Matrix, with an accuracy value of 0.820 (82.0%), precision value of 0.8193 (81.93%), recall value of 0.8200 (82.0%), and a balance value (F-Measure) of 0.8155 (81.55%). The satisfaction assessment from 30 users showed an average score in the "good level" (4.20 ± 0.83). The LINE chatbot developed in this research effectively provides cosmetic advertising information with good accuracy and user satisfaction. It can be improved for greater accuracy and used to provide reliable information to entrepreneurs seeking guidance on cosmetic advertising.