RT Dissertation/Thesis
T1 A general framework for cost-sensitive boosting
A1 Landesa Vazquez, Iago
K1 1203.04 Inteligencia Artificial
K1 1209.04 Teoría y Proceso de decisión
AB Boosting algorithms have been widely used to tackle a plethora of problems. Amongthem, cost-sensitive classification stands out as one of the scenarios in which Boostingis most frequently applied in practice. In the last few years, a lot of approaches havebeen proposed in the literature to provide standard AdaBoost with asymmetric capabilities,each with a different focus. However, for the researcher, these algorithmsshape a confusing heap with diffuse differences and properties, lacking a unifiedframework to jointly compare, classify, analyze and discuss the approaches on a commonbasis.Motivated by the preeminent role of AdaBoost in the Viola-Jones framework forobject detection in images, a markedly asymmetric learning problem, in this thesiswe try to untangle the different Cost-Sensitive AdaBoost alternatives presented inthe literature, demystifying some preconceptions and making novel proposals (Cost-Generalized AdaBoost and AdaBoostDB) with a full theoretical derivation. We try toclassify, analyze, compare and discuss this family of algorithms in order to builda general framework unifying them. Our final goal is, thus, being able to find adefinitive scheme to translate any cost-sensitive learning problem to the AdaBoostframework while shedding light on which algorithm ensures the best performanceand formal guarantees.
YR 2014
FD 2014-07-28
LK http://hdl.handle.net/11093/259
UL http://hdl.handle.net/11093/259
LA eng
DS Investigo
RD 18-ene-2025