Sentiment Analysis Related to Law No. 6 Of 2023 on the Employment Cluster Using the Bidirectional Long Short-Term Memory Algorithm

The enactment of UU Cipta Kerja has triggered varied public responses, particularly concerning employment provisions like fixed-term employment (PKWT), the legalization of outsourcing, unfair severance pay, and ease of layoffs. Social media has become a primary platform for the public to share opinions on issues within the law’s employment cluster. This study employs sentiment analysis using the Bidirectional Long Short-Term Memory (Bi-LSTM) algorithm to understand public sentiment about UU Cipta Kerja and sentiment within its content. Bi-LSTM is chosen for its ability to capture temporal relationships and context in long texts, which aids in handling complex sentiment classification. The findings indicate varied public perceptions: neutral sentiment dominates issues like "PKWT" and "Minimum Wage" on Twitter (X), reflecting uncertainty. Positive sentiment appears around "Outsourcing" and "Minimum Wage" provisions, indicating perceived flexibility. Conversely, negative sentiment dominates issues like "Layoffs" and "Severance," on both social media and in UU Cipta Kerja content, signaling concerns over worker rights. The Bi-LSTM model achieved 70.15% accuracy for the Twitter dataset and 83.22% for the law’s content dataset.