Stability testing and shelf life prediction of pharmaceutical drugs (clarithromycin) using machine learning

Stability testing is essential for predicting the shelf life of pharmaceutical drugs, ensuring they remain effective and safe under various storage conditions. Conventional stability testing methods, including real-time and accelerated studies, require significant time and resources. This paper introduces a machine learning-based framework leveraging linear regression and quadratic modeling to predict the shelf life of pharmaceutical products. By analyzing key factors such as temperature, humidity, light exposure, packaging materials, and chemical composition, the proposed approach achieves enhanced prediction accuracy and reduces the duration of stability studies. Experiments using real-world stability data demonstrate that the model achieves over 85providing pharmaceutical companies with a reliable and efficient alternative to traditional methods. The framework also supports faster decision making and regulatory compliance, potentially accelerating the product development lifecycle.