Predictive modeling for shelf life estimation of sunflower oil blended with oleoresin rosemary (Rosmarinus officinalis L.) and ascorbyl palmitate at low and high temperatures

•Oxidative stability of sunflower oil blended with rosemary and ascorbyl palmitate.•Quantitative study of the chemical composition parameters on oxidative stability.•Development of PLS, Rancimat and unified models for shelf-life prediction.•Unified model estimated the shelf life of sunflower oil blends with minimum error.

The induction period (IP) of the formation of conjugated dienes (IPCDV) at low (60 °C) and IP measured using Rancimat at high (100-130 °C) temperatures were determined as oxidative stability measure (OSM) of sunflower oil (SO) samples blended with oleoresin rosemary (Rosmarinus officinalis L.) and ascorbyl palmitate. The relationship between OSM and compositional parameters (peroxide value, acid value, total polar matter, antioxidant capacity and total added antioxidants) was established using partial least square (PLS) regression. The predicted shelf life at 60 °C (SL60) using the PLS and Rancimat models resulted in the under-prediction by 0.02 and 50.22%, respectively. To overcome the shortcomings of Rancimat model, a unified model was developed using IPCDV values as a function of IP at 100-130 °C which over-predicted the SL60 by 0.08%. The unified model estimated the SL25 with an error of ±3.32% which was comparable to PLS (±2.99) and lesser than Rancimat (±11.22) models.