Table 8. Performance of the algorithm for whooping sound detection using test data.

Case	Diag	Whoop	TP	TN	FP	FN	Sen (%)	Spe (%)	PPV(%)	NPV(%)
1	P	Y	1	528	0	0	100.00	100.00	100.00	100.00
2	P	Y	15	234	2	0	100.00	99.15	88.24	100.00
3	P	N	0	99	0	0	-	100.00	-	100.00
4	P	Y	1	113	0	2	33.33	100.00	100.00	98.26
5	P	Y	4	165	0	0	100.00	100.00	100.00	100.00
6	P	Y	1	565	0	1	50.00	100.00	100.00	99.82
7	P	Y	0	640	0	5	0.00	100.00	-	99.22
8	P	Y	0	229	0	1	0.00	100.00	-	99.57
9	P	Y	5	127	2	1	83.33	98.45	71.43	99.22
10	P	N	0	82	0	0	-	100.00	-	100.00
11	NP	N	0	364	0	0	-	100.00	-	100.00
12	NP	N	0	387	0	0	-	100.00	-	100.00
13	NP	N	0	95	0	0	-	100.00	-	100.00
14	NP	N	0	265	0	0	-	100.00	-	100.00
15	NP	N	0	102	0	0	-	100.00	-	100.00
16	NP	N	0	332	0	0	-	100.00	-	100.00
17	NP	N	0	248	0	0	-	100.00	-	100.00
18	NP	N	0	75	0	0	-	100.00	-	100.00
19	NP	N	0	207	0	0	-	100.00	-	100.00
20	NP	N	0	275	0	0	-	100.00	-	100.00
21	NP	N	0	95	0	0	-	100.00	-	100.00
Total			27	5227	4	10	72.97	99.92	87.10	99.81

Diag—indicates whether recording has pertussis (P) or non-pertussis (NP) diagnosis. Whoop—indicates whether recording has whooping sound; Y—Yes; N- No. TP—true positives; TN—true negatives; FP—false positives; FN—false negatives. Sen—sensitivity; Spe—specificity; PPV—positive predictive value; NPV—negative predictive value.