Table 5.

Average precision, recall and F1-score obtained for 50 most frequent ICD codes with combinations of feature selection methods (FSM) and classification methods (CM). Underlined numbers represent the maximum value in each measure across all models. Numbers in boldface represent the highest values within each machine learning model

		Macro			Micro
CM	FSM	Precision	Recall	F1-score	Precision	Recall	F1-score
DT	None	0.512	0.417	0.453	0.494	0.428	0.454
	FCBF	0.680	0.411	0.479	0.644	0.400	0.470
	IG	0.556	0.411	0.464	0.532	0.419	0.462
	Relief	0.574	0.424	0.476	0.539	0.425	0.467
	Chi-square	0.680	0.411	0.479	0.589	0.440	0.494
	SU	0.632	0.443	0.506	0.603	0.437	0.495
	CFS	0.686	0.414	0.483	0.643	0.422	0.486
	mRMR	0.642	0.456	0.518	0.604	0.452	0.506
NB	None	0.201	0.400	0.254	0.238	0.416	0.293
	FCBF	0.568	0.611	0.572	0.533	0.590	0.543
	IG	0.421	0.547	0.451	0.423	0.547	0.459
	Relief	0.490	0.577	0.495	0.473	0.563	0.487
	Chi-square	0.548	0.588	0.536	0.514	0.576	0.517
	SU	0.546	0.587	0.540	0.519	0.567	0.523
	CFS	0.573	0.603	0.572	0.540	0.579	0.546
	mRMR	0.554	0.608	0.563	0.526	0.581	0.539
LR	None	0.121	0.504	0.189	0.163	0.526	0.241
	FCBF	0.578	0.618	0.577	0.542	0.600	0.551
	IG	0.543	0.561	0.533	0.423	0.552	0.523
	Relief	0.550	0.568	0.535	0.530	0.550	0.521
	Chi-square	0.569	0.601	0.569	0.540	0.579	0.546
	SU	0.567	0.604	0.569	0.541	0.580	0.547
	CFS	0.578	0.617	0.580	0.554	0.585	0.556
	mRMR	0.588	0.612	0.585	0.558	0.586	0.559
SVM	None	0.294	0.502	0.363	0.319	0.487	0.378
	FCBF	0.499	0.659	0.520	0.480	0.632	0.511
	IG	0.543	0.522	0.138	0.524	0.472	0.187
	Relief	0.441	0.674	0.467	0.446	0.620	0.473
	Chi-square	0.506	0.628	0.510	0.486	0.604	0.503
	SU	0.502	0.629	0.514	0.487	0.603	0.507
	CFS	0.484	0.686	0.527	0.480	0.640	0.517
	mRMR	0.479	0.684	0.516	0.475	0.642	0.512
DL	None	0.602	0.408	0.472	0.581	0.417	0.474
	FCBF	0.658	0.366	0.442	0.630	0.370	0.445
	IG	0.629	0.389	0.462	0.608	0.403	0.469
	Relief	0.598	0.341	0.411	0.577	0.346	0.412
	Chi-square	0.622	0.396	0.467	0.603	0.408	0.473
	SU	0.626	0.400	0.471	0.607	0.411	0.477
	CFS	0.637	0.372	0.444	0.616	0.385	0.454
	mRMR	0.640	0.369	0.442	0.618	0.384	0.453