Research on emotional state identification and intervention strategies for preschool students using deep learning techniques

As an advanced machine learning technology, deep learning has significant advantages in the fields of image recognition, speech recognition and natural language processing, etc., and its application in emotion recognition has also received increasing attention. The aim of this paper is to utilize deep learning technology to accurately identify the emotional state of preschool students and develop effective intervention strategies based on it. In this study, a database containing rich emotional information was constructed by collecting data on preschool students' facial expressions, speech and text. Then, deep learning methods are used to train and learn from these data in order to realize the automatic recognition of preschool students' emotional states. Feature extraction and classification of facial expressions by CNN, recognition of emotional features in speech signals by RNN, as well as judgment of emotional tendencies in text data by text sentiment analysis model. Based on the recognition of emotional states, this study further explores the development and implementation of intervention strategies. The experimental results show that the use of deep learning technology for pre-school students' emotional state recognition has high accuracy and reliability. Meanwhile, the intervention strategy based on emotion recognition can effectively improve students' learning experience and learning effect. This study not only enriches the application practice of emotion recognition technology in education, but also provides new ideas and methods for emotion education and intervention in preschool education.