Table 4.

Contingency table test set per patient

Smoking status	Reference
	Ever smoker	Never smoker	No information	Total
Data mining algorithm
Ever smoker	160	19	3	182
Never smoker	2	97	0	99
Not classified	7	3	203	213
Total	169	119	206	494

Ever smoker: precision 87.9% (95% CI 82.1–92.1%), recall 94.7% (95% CI 89.8–97.4%), F1-score 0.91. Never smoker: precision 98.0% (95% CI 92.2–99.7%), recall 82.0% (95% CI 73.1–87.8%), F1-score 0.89.