Performance Analysis of Machine Learning Models for Predicting Order Process Management

The manufacturing industry encounters many challenges in sustaining efficiency, quality, and timely delivery in production processes. An adequate management process is key to overcoming frequent delays, optimizing resource allocation, and following production schedules. Machine learning (ML) has transformed predictive analytics and enabled more accurate predictions of delays and improvement in manufacturing workflows. This study presents a performance analysis of different ML models applied to process management in the manufacturing industry. The data is gathered from different plants such as Bath Tissue, Adhesive Tape, T-shirt, and Toy Manufacturing Plants. IoT devices distributed in these plants collect data from different production processes and transmit it through an IoT gateway to the cloud for storage. Several ML models such as Support Vector Machines (SVM), Decision Tree (DT), K-Nearest Neighbors (KNN) Classifier, Random Forest, Gradient Boosting, and AdaBoost, are used to process the data and predict delays in the production process. The findings show that the SVM has an accuracy of 0.9944 and performed better as compared to the other models in predicting delays, thereby enabling proactive measures to enhance efficiency and productivity. The study highlights the capability of ML to improve process management in the manufacturing industry.