DESIGN OF A DISTRIBUTED MACHINE LEARNING FRAMEWORK FOR LARGE-SCALE SOCIAL MEDIA SENTIMENT ANALYSIS

Abstract

Social media platforms continuously generate massive volumes of unstructured textual data that reflect public opinions and sentiments. Efficient processing of such high-volume data requires distributed and scalable analytical frameworks. This paper presents a machine learning-driven distributed framework for highvolume social media sentiment processing. The proposed framework integrates distributed data ingestion, parallel preprocessing, feature extraction, and machine learning-based sentiment classification. A scalable architecture is designed to support real-time and batch sentiment analytics. Ensemble and deep learning models are employed to enhance classification accuracy. Experimental evaluations are conducted on large-scale social media datasets using realistic workloads. Results demonstrate improved accuracy, precision, and recall compared to conventional sentiment analysis models. The framework effectively handles data scalability while maintaining high analytical performance.