<![CDATA[The rapid growth of digital streaming platforms and global online communities has significantly increased the volume of user generated content, making it difficult for organizations to understand viewer engagement trends in real time. This study develops and evaluates machine learning models for real time sentiment analysis to identify global viewer engagement patterns across large scale digital data streams. Analytical framework is employed by collecting viewer comments and interaction data from multiple online platforms, followed by preprocessing techniques including text normalization, tokenization, and feature extraction. Several machine learning algorithms, including supervised classification models and natural language processing techniques, are trained and evaluated to detect positive, negative, and neutral sentiments in real time. Model performance is assessed using accuracy, precision, recall, and F1 score to determine the most effective approach for large scale sentiment monitoring. The findings demonstrate that optimized machine learning models significantly improve the accuracy and responsiveness of real time sentiment detection, enabling more reliable identification of global viewer engagement trends and behavioral patterns. The integration of automated sentiment analysis also enhances the capability of organizations to process large volumes of streaming textual data efficiently. This research highlights the importance of machine learning driven sentiment analysis systems as strategic tools for understanding global audience engagement, supporting data driven decision making, and improving adaptive content strategies in rapidly evolving digital media environments.]]>
