Quality Evaluation of Classroom Interactive Teaching Based on Software-Defined Fuzzy Network Perception

To achieve the popularization of more innovative teaching modes and promote the rapid development of intelligent classroom teaching technology, it is necessary to design and develop the education and training platform more deeply, integrate with new teaching methods, and achieve an efficient teaching environment. We propose a load-balancing strategy based on deep reinforcement learning considering the problem of classroom terminal load. Based on the embedded classroom teaching platform, in the communication process, this strategy will let the reinforcement learning module model analyze the network. Then, through the advantages of the SDN global view, it obtains the status and information of the whole network, realizing the classroom interaction’s learning process. Optimizing the teaching structure is closer to the key and difficult points of teaching. The fuzzy perception clustering model is designed to improve teaching quality. By the embedded system, we process the data of the fuzzy perception model. Then, we combine fuzzy logic reasoning with the classroom teaching evaluation and utilize the fuzzy neural network to analyze the teaching situation. By training the network to get the optimal migration action decision, it can not only effectively improve the load-balancing effect of multiple controllers, but also reduce the balancing time and enhance the overall performance and stability of the network. Through the experimental analysis, compared with the traditional fuzzy neural network method, the mapping effect of the fuzzy clustering perception classification activation function used in this paper on the output of the normalized layer is verified. The clustering effect is 15.22% higher than that of the PE method and 9.32% higher than that of the SHKwon method. Besides, the improved algorithm improves the prediction rate of the network operation situation awareness. It is especially suitable for embedded systems with limited computing power.