Perceptron Ranking Using Interval Labels with Ramp Loss for Online Ordinal Regression

Table 2

Sensitivity estimation of PRIL, PA-I, and proposed PRIL-RAMP algorithm of three datasets under 0% noise data.

Dataset	Rounds	Method	MAE	RMSE	Spearman’s correlation coefficient	Discard sample	Discard rate	Time (s)
Abalone	T = 1400	PA-I					0.00	0.589
		PRIL					0.00	0.190
		PRIL-RAMP					3.00	0.018
	T = 2800	PA-I					0.00	1.075
		PRIL					0.00	0.388
		PRIL-RAMP					3.46	0.041
	T = 4100	PA-I					0.00	1.471
		PRIL					0.00	0.582
		PRIL-RAMP					3.02	0.062
Parkinsons-uprds	T = 2000	PA-I					0.00	0.449
		PRIL					0.00	0.279
		PRIL-RAMP					8.90	0.302
	T = 4000	PA-I					0.00	0.928
		PRIL					0.00	0.558
		PRIL-RAMP		..			10.05	0.621
	T = 5800	PA-I	..				0.00	1.286
		PRIL					0.00	0.842
		PRIL-RAMP					12.8	0.911
Real estate valuation	T = 140	PA-I					0.00	1.895
		PRIL					0.00	0.078
		PRIL-RAMP					7.14	0.108
	T = 280	PA-I					0.00	3.728
		PRIL					0.00	0.156
		PRIL-RAMP					12.14	0.235
	T = 400	PA-I					0.00	5.954
		PRIL					0.00	0.238
		PRIL-RAMP					13.75	0.318