RACE
发表于 2025-3-30 11:20:17
http://reply.papertrans.cn/87/8633/863269/863269_51.png
SIT
发表于 2025-3-30 15:11:57
S. Dhanalakshmi,M. Poongothaiories into seven subcategories of threat types. In level-1 classification, we outperformed similar systems with a test accuracy of 88.16% on the whole dataset and 88.08% accuracy on test dataset with 30% split, while in level-2 classification of threat tweets (followed by level-1) into its subcatego