Prediction of Standard Minute Value Using Machine Learning in the Garment Industry

The garment industry is a critical component of the global economy, and it has been a major driver of economic growth. The industry faces various challenges, including labour practices, one of which involves the estimation of the standard minute value (SMV). The SMV, representing the time required for a qualified operator to complete a task under standard conditions with appropriate allowances, is often estimated primarily based on engineers' experience. Different individuals may predict the SMV differently. Advancements in technology are expected to standardize SMV prediction and make the production processes more efficient. By knowing SMV accurately in advance, the garment production processes can be improved, thereby reducing cost of producing clothes. In this research, data are collected from a ready-made garment (RMG) industry, with the aim to apply machine learning (ML) based regression models to predict SMV outcomes without depending on industrial engineers. Among the regression models, linear regression (LR), decision tree regression (DTR), and random forest regression (RFR) are chosen for predicting SMV. For the model performance evaluation, mean square error (MSE) and squared correlation coefficient (R2) are calculated. The testing results showed that MSE values fall within 0.004 to 0.006 and R2 range from 0.77 to 0.86, indicating that ML-based regression models are quite accurate in predicting SMV. In addition to providing an efficient method for predicting SMV, this research helps in reducing manpower requirements, enhancing productivity, and minimizing losses.