Long Short-Term Memory-Based Sentiment Analysis Model for Emotion Recognition in Online English Learning Platforms

Traditional emotion assessment methods rely on manual evaluation, which makes it difficult to capture the emotional fluctuations of English learners in online interactions in real time and accurately, thus affecting the evaluation and feedback of interaction quality. To solve this problem, this paper constructs a sentiment analysis model based on long short-term memory (LSTM), which can be integrated into the English learning platform to achieve precise and personalized sentiment analysis and feedback. This paper first collects interaction data from English learning platforms and social networks, constructs a dataset, and preprocesses it for the LSTM model training. By tuning hyperparameters, the model performance is optimized, and finally, a model with sentiment analysis capabilities is obtained. To enhance the portability of the model, this paper modularizes it to facilitate its deployment on online interaction platforms. Experimental results indicate that the model in this paper is better than other models in emotion recognition, with recognition accuracy, recall, precision, and F1-score of 0.94, 0.93, 0.93 and 0.93, respectively, which are all higher than those of the control group models. Additionally, in terms of emotion recognition efficiency, when the text size increases from 2,000 to 10,000, the average processing time of the model in this paper only increases from 49[Formula: see text]ms to 61[Formula: see text]ms, which is lower than that of other models. Therefore, this paper provides theoretical support for the application of sentiment analysis in online education platforms and provides a reference for emotion computing in intelligent education systems.