Table 1.

Selected Publications Using Machine Learning Methods to Identify Clusters or Disease Axes in COPD

Category	PMID	Year	No. of Subjects	No. of Clusters/Axes	Method
Clustering	18248806	2008	415	2 clusters	Fuzzy clustering
	19501190	2009	415	2 clusters	Multidimensional scaling and KHM clustering
	20233420	2010	308	4 clusters	K-means
	20075045	2010	322	4 clusters	Principal components analysis and hierarchical clustering
	21177668	2011	342	3 clusters	K-means
	22154126	2012	102	2 clusters	K-means
	23236428	2012	527	3 clusters	Principal components analysis and hierarchical clustering
	23392440	2013	213	5 clusters	Self-organizing maps
	23613569	2013	1,543	3 clusters	Tree-based clustering
	23536961	2013	157	4 clusters	Factor analysis and k-means
	24563194	2014	8,288	4 clusters	K-means
	25642832	2015	2,164	5 clusters	Factor analysis and random forests clustering
	26773458	2016	364	4 clusters	Network-based stratification
	28943279	2017	9,210	3 clusters	Random forests clustering
	29097431	2017	6,060	5 clusters	Hierarchical clustering
	28637835	2017	17,146	Multiple solutions	Random forests and k-medoids clustering
	29671603	2018	4,606	4 trajectories	Bayesian trajectory modeling
Disease axes	19480658	2009	127	4 disease axes	Principal components analysis
	29771274	2018	8,157	5 disease axes	Factor analysis
	31189730	2019	4,726	6 disease axes	Weighted logistic regression