Cnt Interconnect Performance Prediction Based on Random Forest Regression

As semiconductor process nodes continue to shrink, Cu interconnect technology is approaching its physical limits, making interconnect delay an increasingly critical factor affecting circuit performance. Due to their exceptional electrical conductivity, single-walled carbon nanotube (SWCNT) bundles have emerged as a promising alternative for next-generation interconnects. However, accurately predicting the performance of CNT interconnects before optimizing their parameters remains a significant challenge. Traditional numerical computation methods incur substantial time costs when handling large datasets. To address this issue, this paper proposes a performance prediction model for CNT interconnects based on Random Forest Regression (RFR). Experimental results demonstrate that the proposed model achieves highly accurate predictions and significantly accelerates computation compared to traditional methods.