Frequency-Hopping Transmitter Fingerprint Feature Classification Based on Kernel Collaborative Representation Classifier

Table 2

The classification rate of different treated signals (%).

Number of train sample	50	100	150	200	250	300	350	400	450
KNN
Original	57.6	63.8	67.8	70.4	74.4	78.0	83.4	89.4	94.4
Wavelet	58.4	64.3	68.1	72.6	76.1	80.4	85.9	90.6	95.2
Wave atoms	61.5	66.5	69.6	73.8	77.5	82.6	88.1	92.7	96.9
SVM
Original	63.0	73.4	79.4	81.2	86.4	89.0	91.8	95.4	98.0
Wavelet	65.8	76.1	80.6	83.9	87.3	90.3	92.4	96.4	98.4
Wave atoms	69.4	78.3	81.8	85.2	88.7	91.5	94.6	97.2	98.7
SRC
Original	67.0	76.0	83.4	86.8	91.2	92.6	94.5	96.4	98.2
Wavelet	70.5	78.4	85.2	88.7	92.1	93.4	95.6	96.9	98.4
Wave atoms	73.2	81.7	87.0	90.6	92.9	94.7	96.1	97.0	98.5
CRC
Original	70.4	78.6	84.6	87.4	90.4	92.2	94.0	96.6	98.4
Wavelet	72.8	81.6	85.8	88.3	91.5	93.7	94.9	96.9	98.6
Wave atoms	75.5	83.7	86.7	89.7	92.4	94.1	95.5	97.1	98.9