Аннотация
This paper compares the performance of the Adaptive Neuro-Fuzzy Interface System (ANFIS) and Artificial Neural Network (ANN) by developing small-signal models for Gallium Nitrite (GaN) High electron Mobility Transistor (HEMT). The theoretical background of ANFIS modeling is considered, and implemented in the context of the GaN HEMT. The studied GaN HEMT device has a wide frequency range of 40 GHz. Moreover, the full dataset comprises of 36400 data samples, which enables a through comparison of ANN and ANFIS based models. Through the integration of ANFIS, we unveil a comprehensive approach that captures intricate relationships in small signal responses, allowing for more accurate predictions and an improved understanding of the GaN HEMT behavior. The proposed algorithm has achieved high accuracy in terms of mean square error, mean absolute error, and coefficient of regression ($R^{2}$) for both training and testing data sets. To demonstrate the significance of this paper, the popular feed-forward ANN architecture is utilized with the same internal layer configuration as of ANFIS based models. The results show that ANN based models are more accurate and require less training time as compared to ANFIS based models.