A Data-Driven Framework for Smart Dairy Spray Drying: Multi-Objective Modeling and Optimization Using Literature-Derived Datasets

The dataset was compiled from published spray-drying studies on dairy products through a systematic literature search. Electronic databases including Scopus, Web of Science, and Google Scholar were searched using combinations of the terms "spray drying", "dairy", "milk powder", "process parameters", "yield", "moisture content", and "particle size". Studies were included if they reported at least one of the four target variables alongside the compositional and process parameters required as input features. Studies were excluded if they lacked sufficient detail on feed composition, used non-dairy feedstocks, or reported data in formats that precluded numerical extraction. The screening process followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021) The compiled dataset comprised observations from multiple dairy product types. The number of valid observations differed across targets because not all source studies reported all four quality attributes: product yield (n = 114), final moisture content (n = 185), median particle size (n = 124), and residual thermal ratio (n = 199).