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

Table 3

Sensitivity estimation of PRIL, PA-I, and proposed PRIL-RAMP algorithm of three datasets under 25% noise data.

Dataset	Rounds	Method	MAE	RMSE	Spearman’s correlation coefficient	Discard sample	Discard rate	Time (s)
Abalone	T = 1400	PA-I-25%					0.00	0.035
		PRIL-25%					0.00	0.176
		PRIL-RAMP-25%					21.5	0.225
	T = 2800	PA-I-25%					0.00	0.069
		PRIL-25%					0.00	0.368
		PRIL-RAMP-25%					23.04	0.432
	T = 4100	PA-I-25%					0.00	0.105
		PRIL-25%					0.00	0.552
		PRIL-RAMP-25%					32.04	0.623
Parkinsons-uprds	T = 2000	PA-I-25%					0.00	0.082
		PRIL-25%					0.00	0.297
		PRIL-RAMP-25%					21.25	0.382
	T = 4000	PA-I-25%					0.00	0.296
		PRIL-25%					0.00	0.548
		PRIL-RAMP-25%					19.20	0.786
	T = 5800	PA-I-25%			..		0.00	0.385
		PRIL-25%					0.00	0.854
		PRIL-RAMP-25%					25.10	1.127
Real estate valuation	T = 140	PA-I-25%					0.00	0.421
		PRIL-25%					0.00	0.162
		PRIL-RAMP-25%					22.85	0.208
	T = 280	PA-I-25%					0.00	0.923
		PRIL-25%					0.00	0.348
		PRIL-RAMP-25%		..	..		20.00	0.378
	T = 400	PA-I-25%					0.00	1.263
		PRIL-25%					0.00	0.531
		PRIL-RAMP-25%					30.25	0.579