Enhancing Trustworthy Mining Trend Analytics Through Hybrid Semantic and Advertisement Detection Models

Mining-related social media discussions provide valuable information regarding operational risks, environmental concerns, equipment failures, workforce safety, and industrial market trends. However, large-scale mining discussion streams are increasingly affected by irrelevant promotional content, spam campaigns, and semantically unrelated posts that reduce analytical reliability. This study proposes a hybrid natural language processing framework for trustworthy mining trend analytics through semantic relevance assessment and advertisement detection. The framework combines transformer-based contextual embeddings with supervised machine learning to distinguish operationally relevant mining discussions from promotional and off-topic content. A curated mining-oriented communication dataset containing 12,500 public posts and discussion entries from mining-related channels was analyzed. Semantic relevance was evaluated using transformer embeddings and cosine similarity against predefined mining-topic descriptors, while advertisement detection was performed using TF–IDF features and supervised classifiers including Logistic Regression, Support Vector Machines, Random Forests, and Gradient Boosting. Experimental results demonstrate that the proposed framework significantly improves mining discussion reliability by filtering promotional intrusion and semantic drift. Random Forest and SVM classifiers achieved the best performance with F1-scores exceeding 0.91 in several mining categories. The framework supports intelligent industrial trend monitoring and can be extended to safety analytics, industrial risk monitoring, and operational intelligence systems.

@article{Ayari2026,
  title     = {Enhancing Trustworthy Mining Trend Analytics Through Hybrid Semantic and Advertisement Detection Models},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {17},
  number    = {9},
  year      = {2026},
  publisher = {The Science and Information Organization},
  author    = {Mohamed Ayari and Khulud Alashjaee and Balsam Alzamam and Aryam Alshammari and Milaf Alshammari and Aadil Alshammari},
  doi       = {10.14569/IJACSA.2026.0170981},
  url       = {https://doi.org/10.14569/IJACSA.2026.0170981}
}