Table 13.

Performance results obtained with 2 classes (wheeze vs. other)—training: fixed duration; testing: variable duration.

Classifiers	Accuracy	AUCWheez	F1Wheez	MCCWheez	F1Other	MCCOther
LDA_10MRMR	53.3 ± 0.4	55.4 ± 0.0	63.6 ± 0.5	58.4 ± 0.5	35.2 ± 0.2	30.3 ± 0.2
LDA_100MRMR	53.7 ± 0.6	56.2 ± 1.5	63.8 ± 0.7	58.7 ± 0.7	35.8 ± 0.2	30.9 ± 0.3
LDA_Full	56.6 ± 0.9	56.8 ± 0.7	67.3 ± 1.0	62.5 ± 1.1	35.2 ± 0.2	30.7 ± 0.3
SVMrbf_10MRMR	57.3 ± 1.4	49.1 ± 2.2	69.6 ± 2.1	65.1 ± 2.5	27.2 ± 4.2	23.5 ± 3.6
SVMrbf_100MRMR	57.4 ± 1.7	53.5 ± 1.4	70.3 ± 1.6	65.9 ± 1.9	24.7 ± 2.6	21.2 ± 2.6
SVMrbf_Full	61.2 ± 0.6	57.2 ± 1.1	73.4 ± 0.5	69.5 ± 0.6	28.9 ± 1.8	26.4 ± 1.7
RUSBoost_10MRMR	61.2 ± 0.9	51.7 ± 0.6	74.8 ± 0.7	71.6 ± 0.9	15.4 ± 2.3	14.8 ± 2.3
RUSBoost_100MRMR	62.4 ± 0.5	53.2 ± 0.5	76.0 ± 0.3	73.3 ± 0.5	12.7 ± 1.7	13.4 ± 1.8
RUSBoost_Full	61.3 ± 0.8	52.7 ± 1.9	75.6 ± 0.7	73.0 ± 1.0	5.8 ± 1.3	5.5 ± 1.1
CNN_dualInput	64.1 ± 0.1	50.2 ± 0.1	77.9 ± 0.1	−1.0 ± 0.7	4.7 ± 0.1	−1.0 ± 0.7
CNN_Spectrogram	64.1 ± 0.0	51.2 ± 0.0	77.9 ± 0.0	−1.2 ± 0.2	4.8 ± 0.0	−1.2 ± 0.2
CNN_melSpectrogram	64.0 ± 0.5	50.2 ± 0.2	77.8 ± 0.5	−1.1 ± 1.0	5.1 ± 1.2	−1.1 ± 1.0