A Novel Method for the Separation of Overlapping Pollen Species for Automated Detection and Classification

Table 6

Quantitative classification results.

Technique	Experiment (-fold cross validation)	⁢External quality indicators		⁢Internal quality indicators
		TPR	PPV	HM	FPR
MLP	⁢	0.974	0.975	0.974	0.003
	⁢	0.961	0.962	0.961	0.004
	⁢	0.961	0.962	0.961	0.004
RTF	⁢	0.955	0.955	0.955	0.005
	⁢	0.955	0.956	0.955	0.005
	⁢	0.955	0.955	0.955	0.005
BN	⁢	0.935	0.94	0.935	0.006
	⁢	0.937	0.941	0.937	0.005
	⁢	0.929	0.935	0.929	0.006

BN: Bayesian network; FPR: fallout or false positive rate; HM: harmonic mean; MLP: multilayer perceptron; PPV: precision; TPR: sensitivity, recall, or true positive rate; RTF: random tree forests.