Advantages of Assembly Machine Learning Models for Predicting Employee Productivity in a Garment Manufacturing Company

Goyzueta C.A.R.; De La Cruz J.E.C.; MacHaca W.A.M. · Proceedings of the 2021 IEEE Engineering International Research Conference, EIRCON 2021 (2021) · DOI: 10.1109/eircon52903.2021.9613559

Resumen

The demand for garments throughout the world is very high, it needs to be satisfied, and it is desirable that those who make decisions in the garment industry can predict the productivity performance of work teams in their factories, this research propose assembly models for prediction of employee productivity in a company in the field, the dataset used includes attributes of the manufacturing process and employee productivity. The goal is to show the advantages of assembly algorithms like Bagging Boosting, Gradient Boosting and XGBoost; also grid search cross-validation is used to determine the best model within these methods. The advantage within the assembly models produced in this investigation shows that Gradient Boosting can reduce the mean absolute error (MAE), even over a previous investigation using the same dataset, also assembly models are a robust solution to make predictions in regression cases, even when compared to deep learning models.