Objective: Microbial growth models are typically developed when the objective is to understand the responses of microorganisms when part of the range of conditions studied permits growth to occur. Different methods were used for predicting microbial growth in various conditions. In the present study we have examined a new tool for predictive microbiology; Adaptive Neuro-Fuzzy Inference System (ANFIS). ANFIS is combination of artificial neural networks and fuzzy logic. The proposed neuro-fuzzy model in ANFIS is a multilayer neural network-based fuzzy system. Method & Materials: This eliminates the disadvantage of a normal feedforward multilayer network, which is difficult for an observer to understand or to modify. A dataset consisted of 126 combination of temperature, acetic acid, NaCl concentration, and inoculum level was used in Staphylococcus aureus growth model. In order to develop ANFIS model the data set was divided into three sets of training (63 objects), test set (32 objects) and prediction set (31 objects). The training set was used for constructing the model and the test set for taking care of the overfitting. Results & Conclusion: We optimized the number and type of membership functions using RMSE as a criterion for the test set. The prediction set was used for evaluating the generated ANFIS model as external validation set. In order to further evaluate the model, internal validation was considered using a 10-fold cross-validation technique. The root mean square errors (RMSEs) for the calibration and prediction sets are 0.615 and 0.654, respectively

Keywords