×

Analysis of classification with a reject option. (English) Zbl 1178.68400

Summary: In many classification problems, objects should be rejected when the confidence in their classification is too low. We consider a new classification algorithm with a reject option. Based on the majority vote strategy and plug-in rules, we provide error analysis for this algorithm in ideal and realistic settings, respectively. In addition, some discussions of semi-supervised classification are given to demonstrate our theoretical analysis.

MSC:

68T05 Learning and adaptive systems in artificial intelligence
62J02 General nonlinear regression
Full Text: DOI

References:

[1] DOI: 10.1142/S0219691308002379 · Zbl 1268.42052 · doi:10.1142/S0219691308002379
[2] Bartlett P. L., J. Mach. Learn. Res. 9 pp 1823–
[3] DOI: 10.1016/0167-8655(94)00074-D · doi:10.1016/0167-8655(94)00074-D
[4] DOI: 10.1109/18.556600 · Zbl 0873.68185 · doi:10.1109/18.556600
[5] DOI: 10.1007/s10496-007-0188-4 · Zbl 1141.68557 · doi:10.1007/s10496-007-0188-4
[6] DOI: 10.1007/s11425-007-0100-x · Zbl 1133.68393 · doi:10.1007/s11425-007-0100-x
[7] DOI: 10.1016/j.ins.2009.01.007 · Zbl 1192.68509 · doi:10.1016/j.ins.2009.01.007
[8] DOI: 10.1109/TIT.1970.1054406 · Zbl 0185.47804 · doi:10.1109/TIT.1970.1054406
[9] DOI: 10.1090/S0273-0979-01-00923-5 · Zbl 0983.68162 · doi:10.1090/S0273-0979-01-00923-5
[10] DOI: 10.1017/CBO9780511618796 · Zbl 1274.41001 · doi:10.1017/CBO9780511618796
[11] DOI: 10.1002/cjs.5550340410 · Zbl 1151.62302 · doi:10.1002/cjs.5550340410
[12] Rigollet P., J. Mach. Learn. Res. 8 pp 1369–
[13] DOI: 10.1142/S0219691308002409 · Zbl 1145.68494 · doi:10.1142/S0219691308002409
[14] DOI: 10.1017/CBO9780511812651 · doi:10.1017/CBO9780511812651
[15] DOI: 10.1214/aos/1079120131 · Zbl 1105.62353 · doi:10.1214/aos/1079120131
[16] Vapnik V., Statistical Learning Theory (1998) · Zbl 0935.62007
[17] DOI: 10.1214/07-EJS058 · Zbl 1320.62153 · doi:10.1214/07-EJS058
[18] DOI: 10.1142/S0219691308002616 · Zbl 1152.91768 · doi:10.1142/S0219691308002616
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.