Understanding public opinion on distinct aspects of textual data is likely possible by aspect-based sentiment analysis in a domain-specific fields like government schemes and policies where insightful information is critical to policymaking. This study offers a comprehensive framework leveraging sophisticated models tailored for each ABSA subtask to analyze the Indian government initiatives. The three state-of-the-art components are integrated into ABSA framework that includes BERT + CRF for appropriate aspect extraction, RoBERTa with an attention mechanism for robust aspect category detection, and DistilBERT + GRU with attention for effective and comprehensible sentiment classification. The suggested methodology achieves high accuracy while maintaining computational efficacy by combining the contextual comprehension of BERT variants along with sequence modeling and attention mechanisms. We demonstrate our proposed approach and its effectiveness by extensive trials conducted on a curated dataset of Indian government policies, providing insightful information for policymakers. Furthermore, the model's attention mechanisms emphasize on important textual data contributing to aspect terms, aspect categories, and sentiment polarities in order to provide interpretability. This study promotes data-driven governance, by laying the groundwork for the use of ABSA in policy analysis.