A graph-based soft sensor using feature expansion and multi-hop attention for melt index prediction

Key Points

• The model predicts melt index, showcasing improved prediction accuracy over traditional methods.
• Incorporating multi-hop attention significantly enhances sensor performance, leading to better data interpretation.
• Graph-based modeling allows for effective feature expansion, capturing complex relationships in the data.
• Implications suggest potential for broader applications in industrial process optimization and real-time monitoring.