A BP Neural Network Method for Grade Classification of Loose Damage in Semirigid Pavement Bases

Table 2

Part of the training sample set.

Attributes
Sample	Average absolute value amplitude	Root mean square amplitude	Amplitude standard deviation	Amplitude deviation	Pulse factor	Waveform factor	Crest factor	Kurtosis factor	Kurtosis
1	45.83	178.68	177.65	31656.78	4.33	12.84	1.30 E + 10	55.73	43.04
2	41.38	174.03	177.19	31568.74	4.23	12.92	1.29 E + 10	56.04	44.90
3	38.21	176.94	175.61	31650.92	4.15	12.93	1.30 E + 10	57.22	42.39
4	41.29	176.45	176.54	31425.52	4.28	12.54	1.29 E + 10	54.66	41.58
5	39.49	177.28	179.34	31338.37	3.86	13.13	1.30 E + 10	54.01	40.45
6	42.46	178.08	178.72	31815.35	4.31	12.77	1.30 E + 10	51.66	40.96
7	43.04	174.59	177.12	31508.84	4.13	13.35	1.30 E + 10	54.16	41.12
8	42.05	178.87	175.06	31373.60	3.86	12.56	1.29 E + 10	51.19	43.30
9	39.89	176.52	175.89	31836.04	4.38	12.90	1.29 E + 10	56.91	40.16
10	45.72	177.35	175.16	31386.90	3.97	13.18	1.30 E + 10	54.73	41.29