Automatic Customer Satisfaction Predictions Based on Indonesian Chatbot Text Data

Abstract

Ensuring customer satisfaction when interacting with chatbots is a major challenge in today's digital age. This study created a model to estimate customer satisfaction levels from chatbot conversation data using Bayes' Theorem, which allows user sentiment to be grouped based on probability. This approach involves collecting chatbot log data, preliminary text processing, and applying the Naive Bayes algorithm to build a prediction model. Testing shows an accuracy rate of 85% in identifying positive, neutral, and negative satisfaction categories. These findings prove the model's effectiveness in improving chatbot service responsiveness, enabling its application in real-time systems to optimize customer experience. This research contributes to the fields of artificial intelligence and data analytics, with practical benefits for digital service providers.