Personalized English education system and method based on multimodal sentiment analysis

Abstract

The present application provides a personalized English education system and method based on multimodal sentiment analysis, the system comprising: a multimodal interaction module for receiving and processing students' multimodal data, a sentiment recognition module for performing sentiment analysis on the input multimodal data, a personalization module for evaluating the student's real-time learning status based on the student's historical learning data, and a core brain module for dynamically adjusting interactive feedback based on the sentiment status and real-time learning status. The sentiment recognition module includes sentiment information fusion, which adopts a weighted strategy and adjusts the final output sentiment status through sentiment consistency constraints and conflict correction mechanisms. The present application accurately identifies students' emotions and cognitive states through a sentiment consistency loss function, a conflict correction mechanism, and an out-of-scope control mechanism based on a knowledge graph.

Images