Prediction of Cancer Proteins by Integrating Protein Interaction, Domain Frequency, and Domain Interaction Data Using Machine Learning Algorithms

Table 3

Performance comparison by voting with the majority for five classification types.

Type	Classifiers	ACC	SPE	SEN	1	MCC	PPV	AUC
Bayes	TOP: 3	0.672	0.653	0.695	0.672	0.346	0.609	0.673
	TOP: 5	0.599	0.571	0.660	0.614	0.215	0.410	0.614
Functions	TOP: 5	0.673	0.653	0.700	0.676	0.349	0.600	0.676
	TOP: 3	0.669	0.652	0.692	0.672	0.342	0.606	0.673
Lazy	TOP: 3	0.743	0.837	0.688	0.755	0.504	0.883	0.763
Rules	TOP: 5	0.764	0.825	0.721	0.769	0.537	0.856	0.774
	TOP: 7	0.764	0.826	0.721	0.769	0.537	0.857	0.774
	TOP: 9	0.765	0.817	0.726	0.769	0.534	0.848	0.771
	TOP: 3	0.755	0.838	0.705	0.766	0.528	0.877	0.771
Trees	TOP: 11	0.781	0.831	0.744	0.786	0.567	0.859	0.788
	TOP: 9	0.780	0.820	0.748	0.783	0.563	0.842	0.785
	TOP: 7	0.779	0.816	0.748	0.781	0.561	0.838	0.782
	TOP: 3	0.777	0.811	0.749	0.779	0.558	0.831	0.780
	TOP: 5	0.776	0.811	0.748	0.779	0.556	0.830	0.780