Artificial intelligence-driven sentiment analysis and optimization of movie scripts

This paper introduces a novel artificial intelligence-driven sentiment analysis and optimization approach for movie scripts, aiming to enhance both the accuracy and emotional expression within the script. Utilizing the BERT model, the proposed method automatically classifies the emotions in movie scripts and applies an optimization algorithm to refine emotion fluctuations and transitions. The research focuses on preprocessing the script data, including text cleaning, sentiment labelling, and word segmentation, to ensure standardized input for sentiment analysis. Additionally, the emotion optimization algorithm enhances the accuracy of sentiment analysis results while boosting emotional depth. Cross-validation and hyperparameter tuning guarantee the stability and generalizability of the model. Experimental results demonstrate that this sentiment analysis and optimization method achieves high accuracy across various film script genres and shows significant potential for enhancing script emotional quality in creative industries.