International Research Journal of Education and Technology

The healthcare industry is rapidly evolving, with advancements in Artificial Intelligence (AI) and machine learning playing a crucial role in improving patient care and operational efficiency. However, challenges such as scalability issues, limited predictive accuracy, and complex system implementation remain prevalent in current healthcare prediction models. This work presents an AI-driven approach to health insurance prediction using Graph Neural Networks (GNNs) integrated with cloud computing to address these challenges. The workflow begins with data collection, involving patient demographics, medical history, and wearable sensor data, this collected data is then passed through data preprocessing, where label encoding is applied to categorical variables and outliers are detected using the IQR method, resulting in clean and structured data for analysis. The pre-processed data is used for feature extraction, where statistical measures such as standard deviation and variance are calculated to capture the variability in the data, providing more meaningful input for model training. These extracted features are then used in model development, where GNNs are employed to learn complex relationships between the entities and improve predictive accuracy and scalability. The model is integrated into a cloud-based environment, allowing for seamless deployment and efficient processing of large datasets. The model's performance metrics include an accuracy of 99.44%, precision of 99.32%, sensitivity of 99.12%, specificity of 99.17%, and F-measure of 99.28%. The latency analysis reveals a linear increase, with latency reaching 417 ms for a 150 GB dataset. This work provides an effective and scalable solution for health insurance prediction, utilizing GNNs and cloud computing, which outperforms traditional methods in terms of both accuracy and computational efficiency.

Keywords: Health Insurance Prediction, Deep Learning, Graph Neural Networks (GNNs), Cloud Storage.