Umda omkhulu ummelwane okufutshane
Umda omkhulu okufutshane nommelwane ( LMNN ) [1] ukuhlelwa yi- algorithm yokufunda ngomatshini wokufunda nge-metric . Ifunda ipseudometric eyenzelwe ukuhlelwa kommelwane okufutshane k . I-algorithm isekelwe kwi -semifinite programming, i-sub-class ye -convex optimization .
Injongo yokufunda ephantsi kweliso (ukuhlelwa ngokukodwa) kukufunda umgaqo wesigqibo onokuhlela iimeko zedatha ngokweeklasi ezichazwe kwangaphambili. Umthetho ongummelwane okufutshane uthatha iseti yedatha yoqeqesho yemizekelo ebhaliweyo (okt iiklasi ziyaziwa). Ihlela umzekelo omtsha wedatha kunye neklasi efunyenwe kwivoti yesininzi semizekelo yoqeqesho ekufutshane (ebhaliweyo). Ukusondela kulinganiswa nge- metric echazwe kwangaphambili .
- ↑ Weinberger, K. Q.; Blitzer J. C.; Saul L. K. (2006) "Distance Metric Learning for Large Margin Nearest Neighbor Classification" Advances in Neural Information Processing Systems 18: 1473–1480