Feature selection for medical product demand forecasting with exogeneous variables using Google trend

• Main Author: Koh, Tzi Yong
• Other Authors: Jagath C Rajapakse
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/166059
• Institution: Nanyang Technological University

In an ever-ageing world, the need for enhanced healthcare services has become a paramount global challenge, requiring efficient supply chain management to streamline resource and material management and reduce costs. In healthcare supply chain management, the ability to predict demand is a critical factor, and having a reliable forecasting model is essential for informed decision-making. To this end, we propose a unique approach for forecasting the demand for medical products through the analysis of disease-related keyword trends over time, utilizing data from Google Trends. The results demonstrate that incorporating feature selection methods to select important keywords from Google Trends data can significantly improve the accuracy of forecasting models compared to univariate models. Among the various feature selection techniques tested, Pearson's correlation demonstrated the highest level of effectiveness for the set of external covariates analyzed.