AI-guided optimization of traditional bulgur pilafs: Enhancing sensory and bioactive properties through rsm-pso modeling

This study aimed to enhance the sensory and bioactive properties of pilafs prepared from three geographically indicated bulgur varieties—Siyez, Firik, and Karakilçik—through an AI-guided optimization approach combining Response Surface Methodology (RSM) and Particle Swarm Optimization (PSO). Different bulgur (130–150 g) and water (350–450 mL) ratios were tested to determine optimal formulations. Sensory evaluation revealed that Firik bulgur pilaf achieved the highest overall acceptability (8.49), while Karakilçik bulgur pilaf scored highest in color (7.68) and aroma (8.58), and Siyez bulgur pilaf received the highest taste score (7.50). In terms of bioactive properties, Karakilçik bulgur pilaf showed the highest antioxidant capacity (75.57% DPPH radical scavenging activity), whereas Firik bulgur pilaf had the highest total phenolic (842.39 mg GAE/kg) and flavonoid contents (6.38 mg CE/g). Color analysis indicated that Siyez bulgur pilaf had the lightest color (L=52.18), while Firik pilaf exhibited the most intense red hue (a=8.12) and Karakilçik pilaf the darkest appearance (L=35.42). PSO-based validation confirmed the accuracy of RSM models by reaching global optima within 40 iterations and minimal deviation from experimental values. This is the first study to apply an integrated RSM–PSO modeling approach to traditional bulgur pilafs, enabling the prediction and optimization of their sensory and bioactive characteristics. The results provide a novel framework for enhancing the nutritional value and consumer appeal of heritage cereal-based foods and support the development of standardized, functional bulgur products for the food industry.