The digital transformation of sports has created new opportunities for athletes with disabilities to N Nparticipate, communicate, and express their experiences through online platforms; however, limited attention has been given to how athletes with intellectual disabilities are represented across digital spaces. This study examines their visibility, portrayal, and engagement patterns on digital sports platforms by integrating computational analysis with interpretative insights. Using a large-scale dataset of social media posts related to global sporting events, the research applies sentiment classification and representation analysis to understand public perceptions and narrative framing. To address the limitations of conventional models in capturing contextual and emotional nuances, a Social Group Optimized Long Short-Term Memory (SGO-LSTM) model is proposed, combining swarm-based optimization with sequential deep learning to enhance convergence and classification performance. The experimental results demonstrate that the proposed model outperforms baseline approaches, achieving an accuracy of 91.2%, precision of 90.6%, recall of 89.9%, F1-score of 90.2%, and ROC-AUC of 92.7%. The findings indicate a predominance of positive and inclusive narratives alongside the persistence of stereotypical and biased representations in digital discourse. By linking sentiment outcomes with representation constructs, the study provides a structured framework for understanding how digital narratives influence inclusion, visibility, and bias in sports media, offering valuable insights for policymakers, sports organizations, and advocacy groups working toward equitable digital representation.
Di Jin (Mon,) studied this question.