A Compressive Sensing Model for Speeding Up Text Classification

Table 2

Accuracies of SVM classifier associated with different SRMs on binary and multiclass classification datasets at different subrates.

Subrate R	SRMs
	DCT	FFT	Block DCT	Block WHT	Block Gaussian
Binary classification
0.1	0.6955	0.7220	0.6975	0.6880	0.6930
0.2	0.7185	0.7135	0.7135	0.7200	0.7055
0.3	0.7195	0.7140	0.7285	0.7215	0.7125
0.4	0.7285	0.7190	0.7265	0.7170	0.7185
0.5	0.7235	0.7195	0.7290	0.7270	0.7145
0.6	0.7255	0.7290	0.7265	0.7280	0.7285
Avg.	0.7185	0.7195	0.7203	0.7169	0.7121
Multiclass classification
0.1	0.8590	0.8575	0.8358	0.8444	0.8227
0.2	0.8616	0.8606	0.8651	0.8636	0.8585
0.3	0.8651	0.8737	0.8666	0.8737	0.8606
0.4	0.8686	0.8702	0.8712	0.8747	0.8712
0.5	0.8712	0.8732	0.8767	0.8691	0.8757
0.6	0.8747	0.8782	0.8803	0.8732	0.8762
Avg.	0.8668	0.8689	0.8660	0.8665	0.8609