Predictive Drug Quality Control Using Machine Learning and Big Data

Quality control is a critical aspect of drug manufacturing. Reducing the number of batches that fail to achieve the required critical quality attributes is a crucial factor for the profitability of the drug manufacturer. In the last few decades, several analytical techniques have been developed and employed by manufacturers to track the quality at various stages of production. This paper analyses and validates a strategy that employs Machine Learning (ML) models trained continuously on big data to predict the finished drugs' Critical Quality Attributes (CQAs), at the early stages of production. The proposed strategy is easy to integrate into traditional pharmaceutical manufacturing without critical modifications.