Preference corrections: capturing student and instructor perceptions in educational recommendations

Smart Learning Environments

Table 2 Experimental Results

		U _s,L	U _p,L	FM	NDCG	UtilityLoss	φ	β₁	β₂
Baseline Methods	MF	0.1590	0.1143	0.0787	0.1094	0.2804	N/A	N/A	N/A
	SVR	0.1794	0.1272	0.0845	0.1156	0.2088	N/A	N/A	N/A
	UBRec	0.1812	0.1338	0.0845	0.1264	0.1838	N/A	N/A	N/A
	Rank_p	0.0720	0.2982	0.0274	0.0390	0.4288	N/A	N/A	N/A
Two-Stage Learning	Without Correction	0.1611	0.2387	0.0622	0.0921	0.1925	0.8400	N/A	N/A
	Student Correction	0.2028	0.1348	0.0810	0.1232	0.1541	0.6100	0.9800	N/A
	Instructor Correction	0.2021	0.1392	0.0826	0.1257	0.1439	0.5800	N/A	0.3000
	Combined Correction	0.2029	0.1418	0.0838	0.1262	0.1366	0.5400	0.9800	0.4200
One-Stage Learning	Without Correction	0.1994	0.2506	0.0738	0.1069	0.0665	0.8000	N/A	N/A
	Student Correction	0.2079	0.1339	0.0832	0.1264	0.1371	0.9800	0.98	N/A
	Instructor Correction	0.2079	0.1339	0.0832	0.1264	0.1371	0.9800	N/A	0.9800
	Combined Correction	0.2079	0.1343	0.0832	0.1263	0.1368	0.98	0.98	0.73