Intelligent Speech Recognition Technology in Japanese using Transformer-based Models and Sentiment Analysis

Language processing applications have greatly benefited from the quick creation of technology for intelligent speech recognition, especially area of Japanese speech analysis. In order to improve accuracy and contextual understanding, this research investigates the integration of models based on transformers and sentiment analysis in Japanese speech recognition. The system successfully extracts phonetic variations, emotional tone, and semantic meaning from spoken Japanese by utilizing pre-trained transformer architectures BERT and T5. The suggested approach uses sentiment classification to detect emotions in speech, which helps with automated transcription, language learning. Comparing experimental results to conventional speech recognition models, a significant improvement in accuracy and sentiment detection is shown. The experimental results with an accuracy score of 97.1%, a precision score of 97.2%, a recall score of 98.11%, and an F1 score of 97.1%. shows this study, how intelligent speech recognition technology can support more organic and contextually aware human-computer interactions in Japanese.