A Deep Neuro-Fuzzy Multi-Objective Framework for Quality Optimization of UV-C Treated Microgreen Juices

Abstract

Ultraviolet C treatment is increasingly applied as a non-thermal preservation technique to enhance the safety and quality of plant based beverages. Microgreen juices produced from fenugreek, sunflower, and radish are valued for their high concentrations of antioxidants and bioactive compounds but are susceptible to quality degradation during storage. Accurate prediction and optimization of quality parameters under varying processing conditions is critical for ensuring consumer accept- ability and functional efficacy. This study focuses on quality assessment of ultraviolet C treated microgreen juices using a data driven computational framework. A curated dataset comprising one hundred experimental samples was utilized, incorporating ultraviolet C exposure duration and storage time as input variables. Quality attributes evaluated include sensory parameters such as taste and aroma, microbial stability measured by yeast count, and visual quality indicated by browning index. Multi objective optimization techniques were applied to identify balanced processing conditions that maintain desirable quality across domains. Results highlight the potential of advanced computational approaches in supporting decision making for minimally processed functional beverages. The integration of sen- sory, microbial, and physicochemical metrics provides a holistic perspective on quality retention and supports intelligent control in non-thermal juice preservation.