Production Time Estimation using Machine Learning Techniques for Medium-Sized and Small Companies

Small and medium-sized firms (SMEs) frequently struggle to employ new digital tools including artificial intelligence (AI) along with machine learning (ML) due to a lack of funding and knowledge. These companies are crucial to the expansion of the global economy. Accurately estimating the time required to produce a product is crucial in manufacturing since it influences how satisfied consumers are with their experience resource allocation, and scheduling effectiveness. In SME manufacturing settings, this study suggests a machine learning-driven approach to forecast production time and possible delays. The method combines feature engineering, data preparation, and several machine learning models trained on actual production datasets, such as Artificial neural networks, naive Bayes, random forests, decision trees, and logistic regression are all employed. Standard metrics like accuracy, precision, sensibility, and Fmeasure are used to determine which prediction model is best. Another method for dealing with unbalanced datasets is stratified $\mathbf{k}$-fold cross-validation. The findings demonstrate how Artificial Neural Networks and Random Forests are superior at identifying intricate links in production data, which raises forecasting accuracy. According to the study, SMEs may more successfully implement Industry $\mathbf{4. 0}$ techniques, improve decision-making, and lower operational uncertainty by using ML-based production time estimation. Additionally, the suggested system architecture facilitates scalability and feasible implementation in actual manufacturing environments.

Hali tarjima qilinmagan