Abstract:

A new AdaBoost algorithm for multiclass classification is presented, which is named FSAMME (forward stagewise additive modeling using a multiclass exponential loss function). Further on, the detailed theoretical justification for FSAMME using a novel multiclass exponential loss function and forward stagewise additive modeling is provided. The proposed algorithm only requires the performance of each weak classifier to be better than random guessing (rather than 1/2) and directly solves the multiclass classification question instead of reducing the multiclass classification problem to multiple twoclass problems. The practical applications in UCI repository and aeorengine faulty samples show that the proposed method has higher classification accuracy, which is evidently higher than that of AdaBoost.M1 is slightly higher than that of AdaBoost.MH, meanwhile the proposed method can enormously reduce the computation cost and meet the demand of online quick diagnosis.

Key words: aeroengine,  , fault , diagnosis,  , ensemble , of , classification , methods,  , multiclass , classification , AdaBoost , algorithm,  , forward , stagewise , additive , modeling,  , exponential , loss , function