Regression analysis for supply chain logged data: A simulated case study on shelf life prediction

The paper illustrates that valuable information can be mined from temperature data collected along the perishable food produce supply chain. Three regression techniques: ordinary least square (OLS), principal component regression (PCR) and latent root regression (LRR) have been used to predict remaining shelf life of tropical seafood products. The results show that LRR is the best of the three regression techniques and works well in predicting remaining shelf life for tropical seafood. The results demonstrate the potential usefulness of utilizing automated temperature data collection (e.g. using RFID sensors) to help achieve a challenging business objective-remote real-time prediction of remaining shelf life of chilled foods.