The exponential growth of user-generated content on social media platforms presents both opportunities and challenges in extracting meaningful insights. Sentiment Analysis (SA), a critical component of contextual mining, enables the identification of subjective information embedded within textual data. This article proposes a novel Kernel Extreme Learning Machine-Based Sentiment Analysis of Social Networks Using Improved Dung Beetle Optimization (KELMSASN-IDBO) model, which combines advanced machine learning and nature-inspired optimization techniques to enhance sentiment classification accuracy. The model follows a structured pipeline: initially, raw textual data undergo thorough preprocessing to eliminate noise and standardize content. Subsequently, semantic features are extracted using Bidirectional Encoder Representations from Transformers (BERT) for effective word embedding. The resulting features are then classified using a Kernel Extreme Learning Machine (KELM), known for its high generalization performance and rapid learning speed. To optimize the performance of KELM, an Improved Dung Beetle Optimization (IDBO) algorithm is employed for fine-tuning hyperparameters. Experimental results demonstrate that the proposed KELMSASN-IDBO model outperforms conventional sentiment analysis techniques in terms of accuracy, efficiency, and robustness. The integration of deep contextual embeddings and hybrid optimization makes the proposed model a powerful tool for extracting sentiments from complex and large-scale social network data.
Tamilarasu et al. (Sun,) studied this question.