EnDL-Migrate: An Ensembled Transformer-XGBoost Framework for Analyzing the Rights of Migrant Labour in Construction Industry

Migrant laborers in the construction sector often face challenges related to wage disparities, workplace safety violations, and legal non-compliance. EnDL-Migrate represents an Ensembled Transformer-XGBoost Framework which identifies and predicts labor rights violations in Thiruvalluvar District’s Construction Industry. EnDL-Migrate uses Transformer-based NLP models to handle labor law analysis and legal case reports in combination with structured socioeconomic data analysis through XGBoost-based predictive analytics. An extensive dataset received training while incorporating three interpretability enhancements through categorical encoding and feature scaling alongside legal policy embeddings. Experimental testing establishes that EnDL-Migrate obtains 97.69% accuracy which exceeds baseline performance levels of Logistic Regression (85.32%) and Random Forest (91.47%) and SVM (89.21%). SHAP analysis determines wage disparities, contract enforcement and workplace safety ratings as the most influential factors which affect labor rights. The developed system offers an advanced AI platform which helps monitor legal compliance and generates recommendations alongside providing support tools for decision-making stakeholders. The explainable AI approach of EnDL-Migrate pushes transformative changes to labor rights protection while supporting data-based developments of legal reforms. Future investigations concentrate on broadening the model across various industries and integrating real-time legal updates as well as refining model generalization by using federated learning methods.