Blakeslea trispora is a filamentous fungus known for producing β-carotene, especially in the all-trans form, a powerful antioxidant and precursor to vitamin A. This study aims to identify and evaluate key factors influencing β-carotene and biomass production using the Plackett–Burman design. Seven variables were tested, including concentrations of dairy sludge, soybean meal, yeast extract, incubation time, temperature, and shaker speed. Among these, soybean meal, dairy sludge, and yeast extract were identified as the most significant contributors to enhanced β-carotene yield and biomass accumulation. These substrates not only offer nutritional value but also present a low-cost and sustainable alternative for culture medium formulation. In the second phase, a Gaussian Process Regression (GPR) model was applied to predict β-carotene concentration and biomass yield based on the selected input parameters. The model was validated using k-fold cross-validation, which showed minimal standard deviation in prediction errors and demonstrated strong predictive performance. The Mean Absolute Percentage Error (MAPE) for biomass and β-carotene prediction was 1.82 % and 4.39 %, respectively. These findings support the combined use of experimental design and machine learning techniques for optimizing fungal fermentation processes in β-carotene production.

Keywords