Table 6.

Results of FudanNLP in three classes for 1500 texts by SVM.

Model	Accuracy and F1-Macro of Different Dimensions
	200		300		500		800
	Acc	F1	Acc	F1	Acc	F1	Acc	F1
FW	0.9393	0.9368	0.9387	0.9358	0.9387	0.9465	0.9473	0.9486
P2	0.602	0.5736	0.5647	0.5188	0.604	0.5774	0.6053	0.578
PW	0.9573	0.9569	0.9587	0.9585	0.9573	0.9568	0.9567	0.9561
FP2	0.938	0.9353	0.9393	0.9359	0.9493	0.9451	0.9513	0.9486
FPW	0.964	0.9635	0.9693	0.9685	0.9653	0.9649	0.9653	0.9649
FPC	0.9673	0.967	0.97	0.9699	0.9687	0.9677	0.968	0.9664
Accuracy and F1-Macro of Baselines
Model	Acc				F1
TF–IDF	0.962				0.9613
LDA	0.9406				0.9397
Word2Vec	0.9567				0.9564