Deep learning for sentiment analysis and topic extraction in health insurance

Abstract

Social media has transformed into a vital channel for real-time, unsolicited feedback in healthcare, yet health insurance providers often lack the tools to mine insights from such data. This study proposes a cloud-based system leveraging deep learning for sentiment analysis and topic modeling tailored to the Commercial and Industrial Medical Aid Society (CIMAS) health insurance in Zimbabwe. Using bidirectional encoder representations from transformers (BERT), a convolutional neural network (CNN), a random forest (RF), and autoencoders, the system processes multilingual data from platforms like Twitter and Facebook, identifying customer concerns in real time. Over 15,000 posts were analyzed, with CNN achieving 91.4% accuracy in sentiment classification and BERTopic extracting coherent themes. The system detected issues such as claim delays, app navigation problems, and unreported anomalies. Findings demonstrate that AI can improve service delivery, customer satisfaction, and responsiveness in African insurance contexts.