Abstract

The Pharmaceutical industry uses equipment and machinery in almost all the manufacturing divisions. Equipment breakdown results in significant operating losses, so PdM is an area of emphasis. Ensemble learning models have been promulgated as practical techniques to enhance the prediction accuracy of a system by integrating many learning algorithms. Among those ML solutions, this research focuses on the applicability of ensemble learning models to predict equipment failures in pharma work orders, coupled with a maintenance history for optimal work order scheduling. Different combinations of learning approaches discussed here are bagging, boosting, and stacking, and their effectiveness is demonstrated on external datasets. The findings prove that ensemble models are superior to separate algorithms in lowering downtime and maintenance expenses while improving performance.