Bearing Fault Identification Method under Small Samples and Multiple Working Conditions

Algorithm 1

SAMME.

	Input: M classifiers, training samples, and test samples.
	Output: A strong classifier and the test sample prediction values.
(1)	Initialize the observation weights ;
(2)	for m = 1 to M:
(3)	Fit a classifier to the training data using weights ;
(4)
(5)
	Where K is the total number of sample classifications;
(6)
(7)	Renormalize ;
(8)	end
(9)	Output
	End