The Implementation of Deep Learning Techniques in Developing Conversational Chatbot as The Source of Vaccination Information

Abstract

The Covid-19 pandemic has hit Indonesia for more than 2 years. To overcome Covid-19, Indonesian government implemented a vaccination program with a target of 70% of the population being vaccinated. However, the recorded population that has been vaccinated to reduce the risk of being exposed to Covid-19 is still low. Several studies have stated that information and invitations to vaccines through mass media are considered insufficient to convince the population to vaccinate. Residents who are still unsure and do not even want to vaccinate need really comprehensive information from experts. To answer this problem, a chatbot that can replace experts in explaining everything related to vaccines can be one solution. This is evidenced by a study which states that the interaction between people who have not been vaccinated with a chatbot that explains about vaccination can reduce the level of doubt of the population about the vaccine by up to 20%. The purpose of this research is to build a chatbot using deep learning technique. Meanwhile, the deep learning technique used to build a conversational chatbot is the Multilayer Perceptron Network (MLP). Based on the result of our study, our chatbot can answer 83% questions correctly out of 30 questions.

Sentiment Analysis on Satusehat Application Using Support Vector Machine Method

“…As the fourth most populous country in the world with an uneven distribution, the impact of COVID-19 will undoubtedly be considerable for Indonesia [8], [9]. One of the government's efforts to tackle the spread of COVID-19 is vaccination [10]- [12].…”