Sentiment analysis has emerged as a critical research domain for understanding user opinions from large-scale online platforms. This study proposes a Rejuvenated Particle Swarm Optimization-Based Classifier (RPSOC) designed to address the challenges of classifying sentiments in massive product review datasets. The Amazon Product Review Dataset comprising four domains—Books, DVDs, Electronics, and Kitchen Appliances—has been employed for evaluation. The proposed framework integrates Support Vector Machine with a Gaussian kernel, while RPSOC optimises its parameters through a modified Particle Swarm Optimization strategy that incorporates fitness variance and mean particle distance to prevent premature convergence. MATLAB has been utilised as the primary tool for experimentation and analysis. Comparative results demonstrate that RPSOC consistently outperforms baseline classifiers, achieving 79.17% classification accuracy against an average of 60.2% by existing methods, along with improved precision, Matthews Correlation Coefficient, and F1-score, thereby validating the effectiveness of the proposed approach.